tanbw commited on
Commit
92674ed
1 Parent(s): dbe1964

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +7 -0
  2. CosyVoice-300M-Instruct/campplus.onnx +3 -0
  3. CosyVoice-300M-Instruct/spk2info.pt +3 -0
  4. CosyVoice-300M-SFT/campplus.onnx +3 -0
  5. CosyVoice-300M-SFT/hift.pt +3 -0
  6. CosyVoice-300M-SFT/spk2info.pt +3 -0
  7. CosyVoice-300M/campplus.onnx +3 -0
  8. CosyVoice-300M/spk2info.pt +3 -0
  9. CosyVoice-ttsfrd/resource/festival/Sable.v0_2.dtd +137 -0
  10. CosyVoice-ttsfrd/resource/festival/VCLocalRules +8 -0
  11. CosyVoice-ttsfrd/resource/festival/dicts/cmu/COPYING +35 -0
  12. CosyVoice-ttsfrd/resource/festival/dicts/cmu/allowables.scm +102 -0
  13. CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmu2ft +21 -0
  14. CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmu_lts_rules.scm +0 -0
  15. CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict-0.4.diff +0 -0
  16. CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict-0.4.out +0 -0
  17. CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict-0.4.scm +0 -0
  18. CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict_compile.scm +41 -0
  19. CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict_extensions.scm +574 -0
  20. CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmulex.scm +466 -0
  21. CosyVoice-ttsfrd/resource/festival/dicts/oald/COPYING +35 -0
  22. CosyVoice-ttsfrd/resource/festival/dicts/oald/ali.oald.lex +0 -0
  23. CosyVoice-ttsfrd/resource/festival/dicts/oald/oald_extensions.scm +1704 -0
  24. CosyVoice-ttsfrd/resource/festival/dicts/oald/oald_lts_rules.scm +0 -0
  25. CosyVoice-ttsfrd/resource/festival/dicts/oald/oaldlex.scm +347 -0
  26. CosyVoice-ttsfrd/resource/festival/dicts/wsj.wp39.poslexR +0 -0
  27. CosyVoice-ttsfrd/resource/festival/dicts/wsj.wp39.tri.ngrambin +0 -0
  28. CosyVoice-ttsfrd/resource/festival/etc/email_filter +47 -0
  29. CosyVoice-ttsfrd/resource/festival/multisyn/multisyn.scm +195 -0
  30. CosyVoice-ttsfrd/resource/festival/multisyn/multisyn_pauses.scm +102 -0
  31. CosyVoice-ttsfrd/resource/festival/multisyn/radio_phones_multisyn.scm +136 -0
  32. CosyVoice-ttsfrd/resource/festival/multisyn/send_xwaves.scm +318 -0
  33. CosyVoice-ttsfrd/resource/festival/multisyn/target_cost.scm +410 -0
  34. CosyVoice-ttsfrd/resource/festival/sable-mode.scm +560 -0
  35. CosyVoice-ttsfrd/resource/festival/scfg.scm +62 -0
  36. CosyVoice-ttsfrd/resource/festival/scfg_wsj_wp20.gram +523 -0
  37. CosyVoice-ttsfrd/resource/festival/tobi.scm +338 -0
  38. CosyVoice-ttsfrd/resource/festival/tobi_rules.scm +1002 -0
  39. CosyVoice-ttsfrd/resource/festival/token.scm +809 -0
  40. CosyVoice-ttsfrd/resource/festival/tokenpos.scm +286 -0
  41. CosyVoice-ttsfrd/resource/festival/tts.scm +304 -0
  42. CosyVoice-ttsfrd/resource/festival/unilex_phones.scm +189 -0
  43. CosyVoice-ttsfrd/resource/festival/voices.scm +361 -0
  44. CosyVoice-ttsfrd/resource/festival/voices/english/kal_diphone/COPYING +43 -0
  45. CosyVoice-ttsfrd/resource/festival/voices/english/kal_diphone/festvox/kal_diphone.scm +266 -0
  46. CosyVoice-ttsfrd/resource/festival/voices/english/kal_diphone/festvox/kaldurtreeZ.scm +978 -0
  47. CosyVoice-ttsfrd/resource/festival/voices/english/kal_diphone/group/kallpc16k.group +0 -0
  48. CosyVoice-ttsfrd/resource/festival/voices/english/rab_diphone/COPYING +38 -0
  49. CosyVoice-ttsfrd/resource/festival/voices/english/rab_diphone/festvox/rab_diphone.scm +263 -0
  50. CosyVoice-ttsfrd/resource/festival/voices/english/rab_diphone/group/rablpc16k.group +0 -0
.gitattributes CHANGED
@@ -33,3 +33,10 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ CosyVoice-ttsfrd/ttsfrd-0.3.6-cp38-cp38-linux_x86_64.whl filter=lfs diff=lfs merge=lfs -text
37
+ CosyVoice-ttsfrd/resource/ws/crf_model filter=lfs diff=lfs merge=lfs -text
38
+ CosyVoice-ttsfrd/resource/ws/xiaoyan.dict filter=lfs diff=lfs merge=lfs -text
39
+ CosyVoice-ttsfrd/resource/ws_chhk/xiaoyan.dict filter=lfs diff=lfs merge=lfs -text
40
+ CosyVoice-ttsfrd/resource/ws_chsh/xiaoyan.dict filter=lfs diff=lfs merge=lfs -text
41
+ CosyVoice-ttsfrd/resource/ws_hk/xiaoyan.dict filter=lfs diff=lfs merge=lfs -text
42
+ CosyVoice-ttsfrd/resource/ws_zhsc/crf_model filter=lfs diff=lfs merge=lfs -text
CosyVoice-300M-Instruct/campplus.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6ac6a63997761ae2997373e2ee1c47040854b4b759ea41ec48e4e42df0f4d73
3
+ size 28303423
CosyVoice-300M-Instruct/spk2info.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef7b502d0086fd0ebee5f191446595159813196ee1825b1ee9ea663bc53c5586
3
+ size 30967505
CosyVoice-300M-SFT/campplus.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6ac6a63997761ae2997373e2ee1c47040854b4b759ea41ec48e4e42df0f4d73
3
+ size 28303423
CosyVoice-300M-SFT/hift.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:91e679b6ca1eff71187ffb4f3ab0444935594cdcc20a9bd12afad111ef8d6012
3
+ size 81896716
CosyVoice-300M-SFT/spk2info.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef7b502d0086fd0ebee5f191446595159813196ee1825b1ee9ea663bc53c5586
3
+ size 30967505
CosyVoice-300M/campplus.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6ac6a63997761ae2997373e2ee1c47040854b4b759ea41ec48e4e42df0f4d73
3
+ size 28303423
CosyVoice-300M/spk2info.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef7b502d0086fd0ebee5f191446595159813196ee1825b1ee9ea663bc53c5586
3
+ size 30967505
CosyVoice-ttsfrd/resource/festival/Sable.v0_2.dtd ADDED
@@ -0,0 +1,137 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <!-- <!DOCTYPE sable SYSTEM "Sable.v0_2.dtd" [ -->
2
+ <!--;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;-->
3
+ <!-- -->
4
+ <!-- Centre for Speech Technology Research -->
5
+ <!-- University of Edinburgh, UK -->
6
+ <!-- Copyright (c) 1998 -->
7
+ <!-- All Rights Reserved. -->
8
+ <!-- -->
9
+ <!-- Permission is hereby granted, free of charge, to use and distribute -->
10
+ <!-- this software and its documentation without restriction, including -->
11
+ <!-- without limitation the rights to use, copy, modify, merge, publish, -->
12
+ <!-- distribute, sublicense, and/or sell copies of this work, and to -->
13
+ <!-- permit persons to whom this work is furnished to do so, subject to -->
14
+ <!-- the following conditions: -->
15
+ <!-- 1. The code must retain the above copyright notice, this list of -->
16
+ <!-- conditions and the following disclaimer. -->
17
+ <!-- 2. Any modifications must be clearly marked as such. -->
18
+ <!-- 3. Original authors' names are not deleted. -->
19
+ <!-- 4. The authors' names are not used to endorse or promote products -->
20
+ <!-- derived from this software without specific prior written -->
21
+ <!-- permission. -->
22
+ <!-- -->
23
+ <!-- THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK -->
24
+ <!-- DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING -->
25
+ <!-- ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT -->
26
+ <!-- SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE -->
27
+ <!-- FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES -->
28
+ <!-- WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN -->
29
+ <!-- AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, -->
30
+ <!-- ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF -->
31
+ <!-- THIS SOFTWARE. -->
32
+ <!-- -->
33
+ <!-- -->
34
+ <!-- An early draft of SABLE 0.1 (Jan 8th 1998) -->
35
+ <!-- (Alan W Black [email protected]) -->
36
+ <!-- This not an official DTD just one that allows a mock up of the -->
37
+ <!-- system for Festival-1.2.1 and later -->
38
+ <!-- -->
39
+ <!-- Modified to be XML compliant April 2nd 1998 -->
40
+ <!-- -->
41
+ <!-- Modified for SABLE 0.2 (August 15, 1998) -->
42
+ <!-- (Richard Sproat [email protected]) -->
43
+ <!-- -->
44
+ <!-- Note that in order to use this as both an XML and SGML DTD, we need -->
45
+ <!-- to declare OMITTAG NO in the accompanying sable.decl for SGML. -->
46
+ <!-- This means that end tags are REQUIRED for a sable document to be -->
47
+ <!-- conformant, even when viewed as an SGML document. -->
48
+
49
+ <!ENTITY % baseelements "EMPH |
50
+ BREAK |
51
+ PITCH |
52
+ RATE |
53
+ VOLUME |
54
+ AUDIO |
55
+ ENGINE |
56
+ MARKER |
57
+ PRON |
58
+ SAYAS |
59
+ LANGUAGE |
60
+ SPEAKER |
61
+ DIV">
62
+
63
+ <!ELEMENT SABLE (#PCDATA| %baseelements; )*>
64
+
65
+ <!-- The non-hierarchical ones -->
66
+ <!ELEMENT BREAK EMPTY >
67
+ <!ATTLIST BREAK
68
+ LEVEL (large|medium|small|none|NUMBER) "medium"
69
+ TYPE (quest|excl|period|comma) #IMPLIED
70
+ MSEC CDATA #IMPLIED
71
+ MARK CDATA #IMPLIED>
72
+ <!ELEMENT AUDIO EMPTY >
73
+ <!ATTLIST AUDIO
74
+ SRC CDATA #IMPLIED
75
+ MODE (background|insertion) "insertion"
76
+ LEVEL CDATA #IMPLIED
77
+ MARK CDATA #IMPLIED>
78
+ <!ELEMENT MARKER EMPTY >
79
+ <!ATTLIST MARKER
80
+ MARK CDATA #IMPLIED>
81
+
82
+ <!-- The non-nestable ones -->
83
+
84
+ <!ELEMENT PRON (#PCDATA) >
85
+ <!ATTLIST PRON IPA CDATA #IMPLIED
86
+ SUB CDATA #IMPLIED
87
+ ORIGIN CDATA #IMPLIED
88
+ MARK CDATA #IMPLIED>
89
+ <!ELEMENT SAYAS (#PCDATA) >
90
+ <!ATTLIST SAYAS MODE (literal|date|time|phone|net|postal|currency|math|fraction|measure|ordinal|cardinal|name) #REQUIRED
91
+ MODETYPE (DMY|MDY|YMD|YM|MY|MD|HM|HMS|EMAIL|URL) #IMPLIED
92
+ MARK CDATA #IMPLIED>
93
+
94
+ <!-- The nestable ones -->
95
+
96
+ <!ELEMENT EMPH ( #PCDATA | %baseelements; )*>
97
+ <!ATTLIST EMPH LEVEL (Strong|Moderate|None|Reduced|NUMBER) "moderate"
98
+ MARK CDATA #IMPLIED>
99
+ <!ELEMENT PITCH ( #PCDATA | %baseelements; )*>
100
+ <!ATTLIST PITCH BASE CDATA "0%"
101
+ MIDDLE CDATA "0%"
102
+ RANGE CDATA "0%"
103
+ MARK CDATA #IMPLIED>
104
+ <!ELEMENT RATE ( #PCDATA | %baseelements; )*>
105
+ <!ATTLIST RATE SPEED CDATA "0%"
106
+ MARK CDATA #IMPLIED>
107
+ <!ELEMENT VOLUME ( #PCDATA | %baseelements; )*>
108
+ <!ATTLIST VOLUME LEVEL CDATA "medium"
109
+ MARK CDATA #IMPLIED>
110
+ <!ELEMENT ENGINE ( #PCDATA | %baseelements; )*>
111
+ <!ATTLIST ENGINE ID CDATA #IMPLIED
112
+ DATA CDATA #IMPLIED
113
+ MARK CDATA #IMPLIED>
114
+
115
+ <!ELEMENT LANGUAGE ( #PCDATA | %baseelements; )*>
116
+ <!ATTLIST LANGUAGE ID CDATA #IMPLIED
117
+ CODE CDATA #IMPLIED
118
+ MARK CDATA #IMPLIED>
119
+ <!ELEMENT SPEAKER ( #PCDATA | %baseelements; )*>
120
+ <!ATTLIST SPEAKER GENDER (male|female) #IMPLIED
121
+ AGE (older|middle|younger|teen|child) #IMPLIED
122
+ NAME CDATA #IMPLIED
123
+ MARK CDATA #IMPLIED>
124
+
125
+ <!ELEMENT DIV ( #PCDATA | %baseelements; )*>
126
+ <!ATTLIST DIV TYPE (sentence|paragraph) #IMPLIED
127
+ MARK CDATA #IMPLIED>
128
+
129
+ <!-- Character entities for latin 1 -->
130
+
131
+ <!ENTITY % ISOlat1 PUBLIC
132
+ "-//SABLE//ENTITIES Added Latin 1 for SABLE//EN"
133
+ "sable-latin.ent" >
134
+ %ISOlat1;
135
+
136
+
137
+
CosyVoice-ttsfrd/resource/festival/VCLocalRules ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ SIOD = siod.scm web.scm cstr.scm fringe.scm
2
+
3
+ .copy_from_est: $(SIOD)
4
+
5
+ $(SIOD) :
6
+ @echo 'Copy $@ from EST/lib/siod'
7
+ -del $@
8
+ copy $(EST)\lib\siod\$@ $@
CosyVoice-ttsfrd/resource/festival/dicts/cmu/COPYING ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ This directory contains an American English Lexicon and letter to
3
+ sournd rules based on CMUDICT 0.4. This distribution falls under
4
+ the following copyright. CMUDICT falls under a simialr free licence
5
+ that has no commercial restrictions.
6
+
7
+ Centre for Speech Technology Research
8
+ University of Edinburgh, UK
9
+ Copyright (c) 1996,1997
10
+ All Rights Reserved.
11
+
12
+ Permission is hereby granted, free of charge, to use and distribute
13
+ this software and its documentation without restriction, including
14
+ without limitation the rights to use, copy, modify, merge, publish,
15
+ distribute, sublicense, and/or sell copies of this work, and to
16
+ permit persons to whom this work is furnished to do so, subject to
17
+ the following conditions:
18
+ 1. The code must retain the above copyright notice, this list of
19
+ conditions and the following disclaimer.
20
+ 2. Any modifications must be clearly marked as such.
21
+ 3. Original authors' names are not deleted.
22
+ 4. The authors' names are not used to endorse or promote products
23
+ derived from this software without specific prior written
24
+ permission.
25
+
26
+ THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK
27
+ DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING
28
+ ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT
29
+ SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE
30
+ FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
31
+ WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
32
+ AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
33
+ ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
34
+ THIS SOFTWARE.
35
+
CosyVoice-ttsfrd/resource/festival/dicts/cmu/allowables.scm ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ (require 'lts_build)
2
+ (set! allowables
3
+ '((a _epsilon_ aa aa1 aa0
4
+ ax ax1 ax0
5
+ eh eh1 eh0
6
+ ah ah1 ah0
7
+ ae ae1 ae0
8
+ ey ey1 ey0
9
+ ay ay1 ay0
10
+ er er1 er0
11
+ y-ax0 y-ah1 y-ah0
12
+ aw aw1 aw0
13
+ ao ao1 ao0
14
+ ih ih1 ih0
15
+ w-ax0 w-ah1 w-ah0
16
+ ow ow1 ow0
17
+ w-ey w-ey1 ey0
18
+ iy iy1 iy0)
19
+ (b _epsilon_ b p )
20
+ (c _epsilon_ k ch s sh t-s )
21
+ (d _epsilon_ d t jh)
22
+ (e _epsilon_ ih ih1 ih0
23
+ ax ax1 ax0
24
+ iy iy1 iy0
25
+ er er1 er0
26
+ ax ah1 ah0
27
+ eh eh1 eh0
28
+ ey ey1 ey0
29
+ uw uw1 uw0
30
+ ay ay1 ay0
31
+ ow ow1 ow0
32
+ y-uw y-uw1 y-uw0
33
+ oy oy1 oy0
34
+ aa aa1 aa0)
35
+ (f _epsilon_ f )
36
+ (g _epsilon_ g jh zh k f)
37
+ (h _epsilon_ hh )
38
+ (i _epsilon_ iy iy1 iy0
39
+ ax ax1 ax0
40
+ ih ih1 ih0
41
+ ah ah1 ah0
42
+ ax ah1 ah0
43
+ ay ay1 ay0
44
+ y
45
+ aa aa1 aa0
46
+ ae ae1 ae0
47
+ w-ax0 w-ah1 w-ah0
48
+ eh eh1 eh0
49
+ er er0 er1 )
50
+ (j _epsilon_ jh y hh zh)
51
+ (k _epsilon_ k )
52
+ (l _epsilon_ l ax-l y ax0-l)
53
+ (m _epsilon_ m ax-m m-ax0 ax0-m m-ax0
54
+ m-ae m-ae1 m-ae0
55
+ m-ih m-ih0 )
56
+ (n _epsilon_ n ng n-y)
57
+ (o _epsilon_ ax ax0 ah1 ah0
58
+ ao ao1 ao0
59
+ ow ow1 ow0
60
+ uw uw1 uw0
61
+ er er1 er0
62
+ aa aa1 aa0
63
+ aw aw1 aw0
64
+ oy oy1 oy0
65
+ uh uh1 uh0
66
+ w
67
+ w-ax0 w-ah1 w-ah0
68
+ aa aa1 aa0
69
+ ih ih1 ih0
70
+ ae ae1 ae0)
71
+ (p _epsilon_ p f)
72
+ (q _epsilon_ k )
73
+ (r _epsilon_ r er1 er er0 )
74
+ (s _epsilon_ s sh z zh ch)
75
+ (t _epsilon_ t th sh ch dh d s zh)
76
+ (u _epsilon_
77
+ ax ax0
78
+ ah ah1 ah0
79
+ uw uw1 uw0
80
+ er er1 er0
81
+ uh uh1 uh0
82
+ y-uw y-uw1 y-uw0
83
+ ax-w ah1-w ah0-w
84
+ y-er y-er1 y-er0
85
+ y-ax y-ax0 y-ah1 y-ah0
86
+ w
87
+ ih ih1 ih0
88
+ ao ao1 ao0
89
+ eh eh1 eh0
90
+ y-uh y-uh1 y-uh0 )
91
+ (v _epsilon_ v f)
92
+ (w _epsilon_ w v f)
93
+ (x _epsilon_ k-s g-z ng-z k-sh z g-zh zh)
94
+ (y _epsilon_
95
+ iy iy1 iy0
96
+ ih ih1 ih0
97
+ ay ay1 ay0
98
+ y
99
+ ax ax0 ah1 ah0)
100
+ (z _epsilon_ z t-s zh s)
101
+ (# #)))
102
+
CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmu2ft ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/sh
2
+ #
3
+ # Converts cmulexicon to Festival format
4
+ #
5
+ # usage: cmu2ft cmudict.0.1 cmu_lex.scm
6
+
7
+ #sed 's/er0/er0 r/' | sed 's/er1/er1 r/' | sed 's/er2/er2 r/' |
8
+
9
+ echo >$2
10
+ echo ";; CMUDICT-0.4 Converted to Festival lexicon format" >>$2
11
+ cat $1 | tr "[A-Z]" "[a-z]" | sed 's/ah0/ax/g' |
12
+ tr 2 1 |
13
+ awk '{if ($1 == "##")
14
+ printf(";; %s\n",$0);
15
+ else if ($1 ~ /^[a-z][a-z]*$/)
16
+ { printf("(")
17
+ printf("\"%s\" nil (%s",$1,$2)
18
+ for (i=3; i <= NF; i++)
19
+ printf " %s",$i
20
+ printf "))\n"
21
+ }} ' >> $2
CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmu_lts_rules.scm ADDED
The diff for this file is too large to render. See raw diff
 
CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict-0.4.diff ADDED
The diff for this file is too large to render. See raw diff
 
CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict-0.4.out ADDED
The diff for this file is too large to render. See raw diff
 
CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict-0.4.scm ADDED
The diff for this file is too large to render. See raw diff
 
CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict_compile.scm ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 2008 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Compile the lexicon
35
+ ;;;
36
+
37
+ (load "cmulex.scm")
38
+ (lex.compile "all.scm" "cmudict-0.4.out")
39
+
40
+
41
+
CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmudict_extensions.scm ADDED
@@ -0,0 +1,574 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1998 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Extra items that have sufficient frequency and are pronounced
35
+ ;;; wrongly that are to be added to the compiled version of
36
+ ;;; the CMULEX lexicon
37
+ ;;;
38
+ ;;; This list we derived from various text databases and actually
39
+ ;;; listening to the LTS models
40
+
41
+ ("dont" v (d ow1 n t))
42
+ ("doesnt" v (d ah1 z n t))
43
+ ("youre" nil (y ao1 r))
44
+ ("wasnt" v (w aa1 z n t))
45
+ ("wouldnt" v (w uh1 d n t))
46
+ ("couldnt" v (k uh1 d n t))
47
+ ("shouldnt" v (sh uh1 d n t))
48
+ ("youve" nil (y uw1 v))
49
+ ("werent" nil (w er1 n t))
50
+ ("Lieut" nil (l uw t eh1 n ax n t))
51
+ ("hed" v (hh iy1 d))
52
+ ("youd" v (y uw1 d))
53
+ ("Ayatullah" n (ay1 ax t uh1 l ax))
54
+ ("intifadeh" n (ih1 n t ax f aa1 d ey))
55
+ ("Gaddafi" n (g ax d ae1 f iy))
56
+ ("Muammar" n (m uw ae1 m m aa1 r) )
57
+ ("prizewinning" n (p r ay1 z w ih1 n ax ng) )
58
+ ("Yevgeni" n (y eh v g eh1 n iy) )
59
+ ("HIV" n (ey1 ch ay1 v iy) )
60
+ ("Lemonick" b (l eh m ah1 n ax k) )
61
+ ("sures" n (sh uw1 r z) )
62
+ ("putsch" n (p uh1 t ch) )
63
+ ("oclock" n (ax k l aa1 k) )
64
+ ("Peterzell" n (p iy1 t er z eh l) )
65
+ ("yds" n (y aa1 r d z) )
66
+ ("mujahedin" n (m uw1 hh ax hh eh d iy n) )
67
+ ("Zedong" n (z ey1 d ao ng) )
68
+ ("Ungeheuer" n (ah n g ax hh uw er) )
69
+ ("antiapartheid" n (ae1 n t iy ax p aa1 r t hh ay t ) )
70
+ ("auteur" n (ow t uw1 r) )
71
+ ("surefire" n (sh uw r f ay1 er) )
72
+ ("moviegoer" n (m uw1 v iy g ow er) )
73
+ ("OHara" n (ow hh aa1 r ax) )
74
+ ("NIH" n (eh1 n ay ey1 ch) )
75
+ ("Sotheby" n (s ao1 th ax b iy) )
76
+ ("Dak" n (d aa1 k) )
77
+ ("itll" n (ih1 t ax l) )
78
+ ("Kryuchkov" n (k r uw1 ch k ax v) )
79
+ ("FDIC" n (eh1 f d ax k) )
80
+ ("Caltech" n (k ae1 l t eh1 k) )
81
+ ("VIP" n (v iy1 ay p iy1) )
82
+ ("whove" n (hh uw1 v) )
83
+ ("Seiichi" n (s ey1 iy ch iy) )
84
+ ("Kanise" n (k ae1 n iy z) )
85
+ ("Gup" n (g uh1 p) )
86
+ ("prizewinner" n (p r ay1 z w ih1 n er) )
87
+ ("genocidal" n (jh eh1 n ax s ay1 d ax l) )
88
+ ("eponymous" n (ax p aa1 n ax m ax s) )
89
+ ("Tallinn" n (t aa1 l iy n) )
90
+ ("Lothar" n (l ao th aa1 r) )
91
+ ("Kampuchea" n (k ae m p ax ch iy1 ax) )
92
+ ("Bessmertnykh" n (b eh s m eh1 r t n ax k) )
93
+ ("yd" n (y aa1 r d) )
94
+ ("atavistic" n (ae1 t ax v ih1 s t ax k ) )
95
+ ("Tadzhikistan" n (t ae1 d zh ax k ax s t aa1 n) )
96
+ ("sadomasochistic" n (s ey1 d ax m ae s ax k ih1 s t ax k) )
97
+ ("impudent" n (ih1 m p y ax d ax n t) )
98
+ ("antinuclear" n (ae1 n t ax n uw1 k l iy er) )
99
+ ("Komsomolskaya" n (k ax m s aa1 m ax l s k ay1 ax ) )
100
+ ("proj" n (p r aa1 jh eh k t) )
101
+ ("ROTC" n (aa1 r ow t iy s iy1) )
102
+ ("Jamil" n (jh ax m iy1 l) )
103
+ ("Agha" n (aa1 g ax) )
104
+ ("skewered" n (s k y uw1 er d) )
105
+ ("celibate" n (s eh1 l ax b ax t) )
106
+ ("Bubiyan" n (b uw1 b ax y aa n) )
107
+ ("Bakatin" n (b ae1 k ax t ax n) )
108
+ ("whod" n (hh uw1 d) )
109
+ ("reverent" n (r eh1 v er ax n t) )
110
+ ("occupiers" n (aa1 k y ax p ay1 r z) )
111
+ ("meanspirited" n (m iy1 n s p ih1 r ax t ax d) )
112
+ ("bluesy" n (b l uw1 z iy) )
113
+ ("aboveground" n (ax b ah1 v g r aw n d) )
114
+ ("Yanayev" n (y ae1 n ay eh1 v) )
115
+ ("VOL" n (v ao1 l) )
116
+ ("Teutonic" n (t uw t aa1 n ax k) )
117
+ ("Serrill" n (s eh1 r ax l) )
118
+ ("Branegan" n (b r ae1 n ax g ax n) )
119
+ ("Bjerklie" n (b eh1 r k l iy) )
120
+ ("unerring" n (uh n eh1 r ax ng) )
121
+ ("sugarcane" n (sh uh1 g er k ey n) )
122
+ ("simpleminded" n (s ih1 m p ax l m ay1 n d ax d) )
123
+ ("menaced" n (m eh1 n ax s d) )
124
+ ("kidnapers" n (k ih1 d n ae p er z) )
125
+ ("detonator" n (d eh1 t ax n ey1 t er) )
126
+ ("dachas" n (d aa1 ch ax s) )
127
+ ("cheerily" n (ch iy1 r ax l iy) )
128
+ ("chanteuse" n (sh ae1 n t ow z) )
129
+ ("acidly" n (ae1 s ax d l iy) )
130
+ ("Kirghizia" n (k ax r g ih1 t s iy ax) )
131
+ ("Isozaki" n (iy1 s ax z aa1 k iy) )
132
+ ("Ilyich" n (iy1 l y ax ch) )
133
+ ("Eurocentric" n (y uw1 r ax s eh1 n t r ax k) )
134
+ ("worshiping" n (w er1 sh ax p ax ng) )
135
+ ("villainy" n (v ih1 l ax n iy) )
136
+ ("souffle" n (s uw f l ey1) )
137
+ ("redolent" n (r eh1 d ax l ax n t) )
138
+ ("probusiness" n (p r ow b ih1 z n ax s) )
139
+ ("postgraduate" n (p ow1 s t g r ae1 jh ax w ax t) )
140
+ ("obstinate" n (aa1 b s t ax n ax t) )
141
+ ("hallucination" n (hh ae l uw1 s ax n ey sh ax n) )
142
+ ("goddam" n (g aa1 d ae m) )
143
+ ("exorcism" n (eh1 k s ax r s ih1 z ax m ) )
144
+ ("daringly" n (d ey1 r ax ng l iy) )
145
+ ("backstabbing" n (b ae1 k s t ae1 b ax ng) )
146
+ ("Wenceslas" n (w eh1 n s ax s l ax s) )
147
+ ("Uzis" n (uw1 z iy z) )
148
+ ("Trabant" n (t r ax b aa1 n t) )
149
+ ("Sinead" n (sh ih1 n ey d) )
150
+ ("Nazarbayev" n (n ae1 z er b ay1 ax v) )
151
+ ("Nadia" n (n aa1 d iy ax) )
152
+ ("unquenchable" n (uh n k w eh1 n sh ax b ax l) )
153
+ ("typecast" n (t ay1 p k ae1 s t) )
154
+ ("therell" n (dh eh1 r ax l) )
155
+ ("syrupy" n (s ih1 r ax p iy) )
156
+ ("superagent" n (s uw1 p er ey1 jh eh n t) )
157
+ ("sourly" n (s aw1 er l iy) )
158
+ ("shuddered" n (sh ah1 d er ax d) )
159
+ ("seceding" n (s eh s iy1 d ax ng) )
160
+ ("preternatural" n (p r iy t er n ae1 t ch er ax l) )
161
+ ("postseason" n (p ow1 s t s iy1 s ax n) )
162
+ ("policymaker" n (p aa1 l ax s iy m ey1 k er) )
163
+ ("nosy" n (n ow1 z iy) )
164
+ ("maleness" n (m ey1 l n ax s) )
165
+ ("luscious" n (l ah1 sh ax s) )
166
+ ("gigolo" n (jh ih1 g ax l ow) )
167
+ ("genially" n (jh iy1 n iy ax l iy) )
168
+ ("epiphanies" n (eh p ih1 f ae n iy z) )
169
+ ("dungeon" n (d ah1 n jh ax n) )
170
+ ("dogmas" n (d ao1 g m ax s) )
171
+ ("discomfited" n (d ax s k aa1 m f ax t ax d) )
172
+ ("deadlier" n (d eh1 d l iy er) )
173
+ ("autoworkers" n (ao1 t ow w er k er z) )
174
+ ("Yanqui" n (y ae1 n k iy) )
175
+ ("Wehrmacht" n (v er m aa1 k t) )
176
+ ("Trabants" n (t r ax b aa1 n t z) )
177
+ ("Swahili" n (s w ax hh iy1 l iy) )
178
+ ("Sodom" n (s aa1 d ax m) )
179
+ ("Seurat" n (s uw1 r ax t) )
180
+ ("NAACP" n (eh1 n d ah1 b ax l ey1 s iy p iy1) )
181
+ ("Manichaean" n (m ae1 n ax k iy ax n) )
182
+ ("Khaled" n (k aa1 l eh d) )
183
+ ("Farouk" n (f aa r uw1 k) )
184
+ ("Bendjedid" n (b eh1 n jh ax d ax d) )
185
+ ("Beetlejuice" n (b iy1 t ax l jh uw s) )
186
+ ("whir" n (w ih1 r) )
187
+ ("tragicomedy" n (t r ae1 jh ax k aa1 m ax d iy) )
188
+ ("touchingly" n (t ah1 ch ax ng l iy) )
189
+ ("thered" n (dh eh1 r ax d) )
190
+ ("tangy" n (t ae1 ng iy) )
191
+ ("synonyms" n (s ih1 n ax n ax m z) )
192
+ ("subspecies" n (s ah1 b s p iy sh iy z) )
193
+ ("subjugated" n (s ax b jh ax g ey1 t ax d) )
194
+ ("stunner" n (s t ah1 n er) )
195
+ ("springy" n (s p r ih1 ng iy) )
196
+ ("sorrows" n (s aa1 r ow z) )
197
+ ("sequins" n (s iy1 k w ax n z) )
198
+ ("schoolgirl" n (s k uw1 l g er l) )
199
+ ("schoolchild" n (s k uw1 l ch ay l d) )
200
+ ("scavenging" n (s k ae1 v ax n jh ax ng) )
201
+ ("rampaging" n (r ae1 m p ey jh ax ng ) )
202
+ ("radiance" n (r ey1 d y ax n s) )
203
+ ("pussycat" n (p uh1 s iy k ae t) )
204
+ ("paparazzi" n (p ae1 p er ae t s iy) )
205
+ ("ovaries" n (ow1 v er iy z) )
206
+ ("languorous" n (l ae1 ng ao r ax s) )
207
+ ("incorruptible" n (ih1 n k er ah1 p t ax b ax l) )
208
+ ("hypnotically" n (hh ih1 p n aa1 t ax k ax l iy) )
209
+ ("grouchy" n (g r aw1 ch iy) )
210
+ ("genitalia" n (jh eh1 n ax t ey1 l y ax) )
211
+ ("fearlessly" n (f iy1 r l ax s l iy) )
212
+ ("divinely" n (d ih1 v ay n l iy) )
213
+ ("bluefish" n (b l uw1 f ax sh) )
214
+ ("bayonets" n (b ey1 y ax n eh t s) )
215
+ ("autoworker" n (ao1 t ow w er k er) )
216
+ ("antihero" n (ae1 n t ax hh iy1 r ow) )
217
+ ("Yel" n (y eh1 l) )
218
+ ("Talmudic" n (t ae1 l m ax d ax k) )
219
+ ("Staci" n (s t aa1 z iy) )
220
+ ("Sese" n (s eh1 s iy1 ) )
221
+ ("Senegalese" n (s ax n ax g ao l iy1 z) )
222
+ ("Quezon" n (k w eh1 z ax n) )
223
+ ("Pesci" n (p eh1 s k iy) )
224
+ ("Mengele" n (m eh1 n g ax l ey) )
225
+ ("Lagerfeld" n (l ae1 g er f eh l d) )
226
+ ("Kerouac" n (k eh1 r ow ae k) )
227
+ ("Kamisar" n (k ae1 m ax s aa r) )
228
+ ("Janata" n (jh ax n ae1 t ax) )
229
+ ("Hequet" n (hh eh1 k ax t) )
230
+ ("Friedan" n (f r iy1 d ax n) )
231
+ ("Dubrovnik" n (d ax b r ah1 v n ax k) )
232
+ ("Candide" n (k ae n d iy1 d) )
233
+ ("Bubba" n (b ah1 b ax) )
234
+ ("AWOL" n (ey1 w ao l) )
235
+ ("wouldve" n (w uh1 d ax v) )
236
+ ("wisecrack" n (w ay1 z k r ae1 k) )
237
+ ("voce" n (v ao1 k ax) )
238
+ ("uproarious" n (ax p r ao1 r iy ax s) )
239
+ ("unearthing" n (ax n er1 th ax ng) )
240
+ ("undoes" n (ax n d ah1 z) )
241
+ ("underappreciated" n (ah1 n d er ax p r iy1 sh iy ey1 t ax d) )
242
+ ("tousled" n (t ao1 s ax l d) )
243
+ ("thatll" n (dh ae t ax l) )
244
+ ("teasingly" n (t iy1 z ax ng l iy) )
245
+ ("strafed" n (s t r ey1 f d) )
246
+ ("stateless" n (s t ey1 t l ax s) )
247
+ ("stagings" n (s t ey1 jh ax ng z) )
248
+ ("reveres" n (r eh1 v ax r iy z) )
249
+ ("retells" n (r iy t eh1 l z) )
250
+ ("provocateur" n (p r ax v ao k ax t uh1 r) )
251
+ ("prenuptial" n (p r iy n ah1 p sh ax l) )
252
+ ("pinup" n (p ih1 n ah1 p) )
253
+ ("paterfamilias" n (p ae1 t er f ax m iy1 l iy ax z) )
254
+ ("paroxysm" n (p ax r aa1 k s ax z ax m) )
255
+ ("parodying" n (p ae1 r ax d iy ax ng) )
256
+ ("merriment" n (m eh1 r ax m ax n t) )
257
+ ("menstruation" n (m ax n s t r ey1 sh ax n) )
258
+ ("indiscretion" n (ih1 n d ax s k r eh1 sh ax n) )
259
+ ("guesthouse" n (g eh1 s t hh aw s) )
260
+ ("fratricidal" n (f r ae t r ax s ay1 d ax l) )
261
+ ("fiendishly" n (f iy1 n d ax sh l iy) )
262
+ ("enthuses" n (eh n th uw1 z ax z) )
263
+ ("encapsulates" n (eh n k ae1 p s ax l ey1 t s) )
264
+ ("discomfiting" n (d ax s k aa1 m f ax t ax ng) )
265
+ ("cul" n (k uh1 l) )
266
+ ("cheekbones" n (ch iy1 k b ow n z) )
267
+ ("bizarrely" n (b ax z aa1 r l iy) )
268
+ ("bestiality" n (b iy1 s t iy ae1 l ax t iy) )
269
+ ("audiotapes" n (aa1 d iy ow t ey1 p s) )
270
+ ("acolyte" n (ae1 k ax l ay t) )
271
+ ("achingly" n (ey1 k ax ng l iy) )
272
+ ("Wayans" n (w ey1 y ax n z) )
273
+ ("Urquart" n (ae1 r k er t) )
274
+ ("Ural" n (y uw1 r ax l) )
275
+ ("Sindayen" n (s ih1 n d ax y ax n) )
276
+ ("Seuss" n (s uw1 s) )
277
+ ("Plexiglas" n (p l eh1 k s iy g l ae s) )
278
+ ("Overbye" n (ow1 v er b ay) )
279
+ ("Nostradamus" n (n aa1 s t r ax d ae1 m ax s) )
280
+ ("Navajos" n (n ae1 v ax hh ow z) )
281
+ ("Massoud" n (m ax s uw1 d) )
282
+ ("Kales" n (k ey1 l z) )
283
+ ("Gio" n (jh iy1 ow) )
284
+ ("Gianfranco" n (jh iy1 ax n f r ae n k ow) )
285
+ ("Gaultier" n (g ao1 l t iy ey) )
286
+ ("Foucault" n (f ow1 k ow l t) )
287
+ ("FSLIC" n (eh1 f s l ax k) )
288
+ ("Disneyesque" n (d ih1 z n ey eh1 s k) )
289
+ ("Corleone" n (k aa1 r l iy ow1 n iy) )
290
+ ("Arnaz" n (aa1 r n ax z) )
291
+ ("Almash" n (ae1 l m ax sh) )
292
+ ("yous" n (y uw1 z) )
293
+ ("whooshing" n (w uw1 sh ax ng) )
294
+ ("volunteerism" n (v ao1 l ax n t iy1 r ax z ax m) )
295
+ ("upended" n (ah1 p eh1 n d ax d) )
296
+ ("unacknowledged" n (ah1 n ax k n aa1 l ax jh d) )
297
+ ("thrillingly" n (th r ih1 l ax g l iy) )
298
+ ("tackiness" n (t ae1 k ax n ax s) )
299
+ ("supernal" n (s uw p eh1 r n ax l) )
300
+ ("superheroes" n (s uw1 p er hh iy1 r ow z) )
301
+ ("superego" n (s uw1 p er iy1 g ow) )
302
+ ("straitlaced" n (s t r ey1 t l ey s d) )
303
+ ("stateroom" n (s t ey1 t r uw m) )
304
+ ("sorcerer" n (s ao1 r s er er ) )
305
+ ("skewering" n (s k y uw1 er ax ng) )
306
+ ("semiprivate" n (s eh1 m ay p r ay1 v ax t) )
307
+ ("saviors" n (s ey1 v y er z) )
308
+ ("roguish" n (r ow1 g ax sh) )
309
+ ("rereleased" n (r iy1 r ax l iy1 s t) )
310
+ ("repressing" n (r ax p r eh1 s ax ng) )
311
+ ("puppeteers" n (p ah1 p ax t iy1 r z) )
312
+ ("provocateurs" n (p r ax v ao k ax t uh1 r z) )
313
+ ("prodemocracy" n (p r ow1 d ax m aa1 k r ax s iy) )
314
+ ("prizewinners" n (p r ay1 z w ih1 n er z) )
315
+ ("presummit" n (p r iy1 s ah1 m ax t) )
316
+ ("pregame" n (p r iy1 g ey1 m) )
317
+ ("preadolescent" n (p r iy1 ae1 d ax l eh s ax n t) )
318
+ ("postindustrial" n (p ow1 s t ax n d ah1 s t r y ax l) )
319
+ ("playgoers" n (p l ey1 g ow1 er z) )
320
+ ("pastiches" n (p ax s t iy1 ch ax z) )
321
+ ("paramour" n (p ae1 r ax m ao r) )
322
+ ("outdoorsman" n (aw1 t d ao r z m ax n) )
323
+ ("ordeals" n (ao r d iy1 l z) )
324
+ ("nuisances" n (n uw1 s ax n s ax z) )
325
+ ("nonmusical" n (n aa1 n m y uh1 z ax k ax l) )
326
+ ("noncombatant" n (n aa1 n k aa1 m b ax t ax n t) )
327
+ ("neurosurgery" n (n y uw1 r ax s er1 jh er iy) )
328
+ ("narcotraficantes" n (n ae1 r k ax t r ae1 f ax k ae1 n t iy z) )
329
+ ("mischievously" n (m ax s ch iy1 v y ax s l iy) )
330
+ ("menaces" n (m eh1 n ax s ax z) )
331
+ ("manifestos" n (m ae1 n ax f eh1 s t ow s) )
332
+ ("magnanimity" n (m ae1 g n ax n ih1 m ax t iy) )
333
+ ("limply" n (l ih1 m p l iy) )
334
+ ("knifing" n (n ay1 f ax ng) )
335
+ ("inveighed" n (ax n v ey1 d) )
336
+ ("inelegant" n (ax n eh1 l ax g ax n t) )
337
+ ("highflyers" n (hh ay1 f l ay1 er z) )
338
+ ("groped" n (g r ow1 p d) )
339
+ ("filmgoers" n (f ih1 l m g ow1 er z) )
340
+ ("eunuchs" n (y uh1 n ax k s) )
341
+ ("ennobling" n (eh n ow1 b ax l ax ng) )
342
+ ("doyenne" n (d ao y eh1 n) )
343
+ ("docudramas" n (d ao k y ax d r ae1 m ax z) )
344
+ ("detours" n (d iy t uh1 r z) )
345
+ ("desecrated" n (d eh1 s ax k r ey1 t ax d) )
346
+ ("decoded" n (d ax k ow1 d ax d) )
347
+ ("dOrsay" n (d ao1 r s iy) )
348
+ ("cynosure" n (s ay1 n ax sh uh r) )
349
+ ("curmudgeonly" n (k er m uh1 d jh ax n l iy) )
350
+ ("cmon" n (k ax m aa1 n) )
351
+ ("cellblock" n (s eh1 l b l aa k) )
352
+ ("carousing" n (k ax r aw1 z ax ng))
353
+ ("blitzing" n (b l ih1 t z ax ng) )
354
+ ("bejeweled" n (b ax jh uw1 ax l d) )
355
+ ("bedevils" n (b ax d eh1 v ax l z) )
356
+ ("bareknuckle" n (b eh1 r n ah1 k ax l) )
357
+ ("awestruck" n (ao1 s t r ah k) )
358
+ ("automatons" n (ao t aa1 m ax t ax n z) )
359
+ ("anteroom" n (ae1 n t ax r uw m) )
360
+ ("anchorwoman" n (ae1 ng k aa r w uh1 m ax n) )
361
+ ("Yahweh" n (y ae1 w ey) )
362
+ ("Vartoogian" n (v er t ow1 g iy ax n) )
363
+ ("Uttar" n (uw1 t er) )
364
+ ("Tsufura" n (t s uw f uw1 r ax) )
365
+ ("Sousuke" n (s ow s uh1 k ey) )
366
+ ("Segelken" n (s eh1 g ax l k ax n) )
367
+ ("Scissorhands" n (s ih1 s er hh ae1 n d z) )
368
+ ("Presbyterians" n (p r eh s b ax t iy1 r iy ax n z) )
369
+ ("Pharaoh" n (f ey1 r ow) )
370
+ ("Nachtwey" n (n ae1 k t w ey) )
371
+ ("Medecins" n (m eh d ax s ae1 n z) )
372
+ ("Manhattanites" n (m ae n hh ae1 t ax n ay1 t s) )
373
+ ("Madama" n (m ax d ae1 m ax) )
374
+ ("Ljubljana" n (l uh1 b l y ae1 n ae) )
375
+ ("Knute" n (k ax n uw1 t) )
376
+ ("Hangin" n (hh ae ng ax n) )
377
+ ("Halaki" n (hh ax l ae1 k iy) )
378
+ ("Gordievsky" n (g ao r d ax eh1 v s k iy) )
379
+ ("Geremek" n (g eh1 r ax m eh k) )
380
+ ("Ehud" n (ey1 hh uw1 d) )
381
+ ("Dieu" n (d y uw1) )
382
+ ("Dachau" n (d ae1 k aw) )
383
+ ("Chaim" n (ch ey1 m) )
384
+ ("Bnai" n (b ax n ay1) )
385
+ ("Antichrist" n (ae1 n t iy k r ay1 s t) )
386
+ ("Afrocentric" n (ae1 f r ax s eh1 n t r ax k) )
387
+ ("wombs" n (w uw1 m z) )
388
+ ("windblown" n (w ih1 n d b l ow n) )
389
+ ("vibrates" v (v ay1 b r ey t s) )
390
+ ("vertebrae" n (v er1 t ax b r ey) )
391
+ ("underpants" n (ah1 n d er p ae1 n t s) )
392
+ ("tuxedoes" n (t ax k s iy1 d ow z) )
393
+ ("tumbleweed" n (t ah1 m b ax l w iy d) )
394
+ ("tryin" n (t r ay1 ax n) )
395
+ ("torturous" n (t ao1 r t y ax s) )
396
+ ("superrich" n (s uw1 p er r ih1 ch) )
397
+ ("subliminally" n (s ax b l ih1 m ax n ax l iy) )
398
+ ("subjugate" n (s ax b jh ax g ey1 t) )
399
+ ("stupefying" n (s t uw1 p ax f ay ax ng) )
400
+ ("stomachache" n (s t ah1 m ax k ey1 k) )
401
+ ("sophistry" n (s ow1 f ax s t r iy) )
402
+ ("sociopath" n (s ow1 sh y ow p ae1 th) )
403
+ ("snugly" n (s n ah1 g l iy) )
404
+ ("serenades" n (s eh1 r ax n ey1 d z) )
405
+ ("sepulchral" n (s ax p uh1 l k r ax l) )
406
+ ("satirize" n (s ae1 t ax r ay z) )
407
+ ("satiny" n (s ae1 t ax n y ) )
408
+ ("runaways" n (r ah1 n ax w ey z) )
409
+ ("rivulets" n (r ih1 v ax l ax t s) )
410
+ ("ricocheting" n (r ih1 k ax sh ey ax ng) )
411
+ ("reimagine" n (r iy ax m ae1 jh ax n) )
412
+ ("refusenik" n (r ax f y uw1 z n ax k) )
413
+ ("ravings" n (r ey1 v ax ng z) )
414
+ ("radicchio" n (r ax d ih1 k iy ow) )
415
+ ("preseason" n (p r iy1 s iy1 s ax n) )
416
+ ("premed" n (p r iy1 m eh1 d) )
417
+ ("preinvasion" n (p r iy1 ax n v ey1 sh ax n) )
418
+ ("precancerous" n (p r iy k ae1 n s er ax s) )
419
+ ("postmenopausal" n (p ow1 s t m eh n ax p ao1 s ax l) )
420
+ ("pinups" n (p ih1 n ah1 p s) )
421
+ ("passerby" n (p ae1 s er b ay) )
422
+ ("parroted" n (p ae1 r ax t ax d) )
423
+ ("paranormal" n (p ae1 r ax n ao1 r m ax l) )
424
+ ("owlish" n (aw1 l ax sh) )
425
+ ("overhyped" n (ow1 v er hh ay1 p t) )
426
+ ("noire" n (n ax w ae1 r) )
427
+ ("motorboats" n (m ow1 t er b ow t s) )
428
+ ("miscreant" n (m ih1 s k r iy ax n t) )
429
+ ("megalomaniacal" n (m eh1 g l ax m ey1 n iy ax k ax l) )
430
+ ("marquees" n (m aa1 r k iy z) )
431
+ ("lapels" n (l ax p eh1 l z) )
432
+ ("kowtowing" n (k aw1 t aw ax ng) )
433
+ ("jeremiads" n (jh eh1 r ax m ay1 ax d z) )
434
+ ("interweaves" n (ih1 n t er w iy1 v z) )
435
+ ("insuperable" n (ax n s uw1 p er ax b ax l) )
436
+ ("insanely" n (ax n s ey1 n l iy) )
437
+ ("inflames" n (ax n f l ey1 m z) )
438
+ ("infest" n (ax n f eh1 s t) )
439
+ ("inebriated" n (ax n iy1 b r ax ey1 t ax d) )
440
+ ("incompletely" n (ih1 n k ax m p l iy1 t l iy) )
441
+ ("impersonate" n (ax m p eh1 r s ax n ey t) )
442
+ ("hysterectomies" n (hh ih1 s t er eh k t ax m iy z) )
443
+ ("hypothalamus" n (hh ay1 p ax th ae1 l ax m ax s) )
444
+ ("hucksterism" n (hh ah1 k s t er ax z ax m) )
445
+ ("hotheads" n (hh aa1 t hh eh1 d z) )
446
+ ("hotheaded" n (hh aa1 t hh eh1 d ax d) )
447
+ ("heavenward" n (hh eh1 v ax n w er d) )
448
+ ("hauteur" n (hh ow t uh1 r) )
449
+ ("gunplay" n (g ah1 n p l ey ) )
450
+ ("griddle" n (g r ih1 d ax l) )
451
+ ("gewgaws" n (g iy1 g ao z) )
452
+ ("gaiety" n (g ey1 ax t iy) )
453
+ ("fuchsia" n (f y uw1 sh ax) )
454
+ ("fixers" n (f ih1 k s er z) )
455
+ ("fidgety" n (f ih1 d jh ax t iy) )
456
+ ("farewells" n (f eh1 r w eh1 l z) )
457
+ ("exoticism" n (eh g z aa1 t ax k ax z ax m) )
458
+ ("enshrines" n (ax n sh r ay1 n z) )
459
+ ("dyslexia" n (d ax s l eh1 k s y ax) )
460
+ ("dynamiting" n (d ay1 n ax m ay t ax ng) )
461
+ ("drumbeats" n (d r ah1 m b iy t s) )
462
+ ("dopers" n (d ow1 p er z) )
463
+ ("doin" n (d uw1 ax n) )
464
+ ("distantly" n (d ih1 s t ax n t l iy) )
465
+ ("dissolute" n (d ih1 s ax l uw t) )
466
+ ("detests" n (d ax t eh1 s t s) )
467
+ ("despots" n (d eh1 s p ax t s) )
468
+ ("denizen" n (d eh1 n ax z ax n) )
469
+ ("decaf" n (d iy1 k ae1 f) )
470
+ ("dIsere" n (d iy s eh1 r iy) )
471
+ ("clanged" n (k l ae1 ng d ) )
472
+ ("bewitching" n (b ax w ih1 ch ax ng) )
473
+ ("barest" n (b eh1 r ax s t ) )
474
+ ("banjoist" n (b ae1 n jh ow ax s t) )
475
+ ("balsamic" n (b ax l s ae1 m ax k) )
476
+ ("auteurs" n (ow t uw1 r z) )
477
+ ("aswarm" n (ae z w er1 m) )
478
+ ("arrivistes" n (ae1 r ax v iy1 s t ax z) )
479
+ ("apparat" n (ae1 p ax r ae t) )
480
+ ("amorality" n (ey1 m er ae1 l ax t iy) )
481
+
482
+ ;;; So what about UK words ?
483
+ ("cheque" n (ch eh1 k) )
484
+ ("cheques" n (ch eh1 k s) )
485
+ ("chequing" n (ch eh1 k ax ng) )
486
+ ("honour" n (aa1 n er) )
487
+ ("honouring" n (aa1 n er ax ng) )
488
+ ("honours" n (aa1 n er z) )
489
+ ("honourable" n (aa1 n er ax b ax l) )
490
+ ("honoured" n (aa1 n er d) )
491
+ ("favour" n (f ey1 v er) )
492
+ ("favours" n (f ey1 v er z) )
493
+ ("favoured" n (f ey1 v er d) )
494
+ ("favourite" n (f ey1 v er ih t) )
495
+ ("favourites" n (f ey1 v er ih t s) )
496
+ ("favouring" n (f ey1 v er ax ng) )
497
+ ("favourable" n (f ey1 v er ax b ax l) )
498
+ ("favourably" n (f ey1 v er ax b l iy) )
499
+ ("labour" n (l ey1 b er) )
500
+ ("labours" n (l ey1 b er z) )
501
+ ("laboured" n (l ey1 b er d) )
502
+ ("labouring" n (l ey1 b er ax ng) )
503
+ ("colour" n (k ah1 l er) )
504
+ ("colouring" n (k ah1 l er ax ng) )
505
+ ("colours" n (k ah1 l er z) )
506
+ ("colourful" n (k ah1 l er f ax l) )
507
+ ("colourfully" n (k ah1 l er f ax l iy) )
508
+
509
+ ;; Added July 2004
510
+ ("tanja" n (t ae1 n y ah))
511
+ ("waibel" n (w ay1 b ax l))
512
+ ("eskenazi" n (eh s k ax n aa1 z iy))
513
+ ("carbonell" nil (k aa1 r b ow n eh1 l))
514
+ ("metze" nil (m eh1 t z ax))
515
+ ("toth" nil (t ow1 th))
516
+ ("adele" nil (ax d eh1 l))
517
+ ("weitz" nil (w ay1 t s))
518
+ ("shamos" nil (sh ey1 m ao s))
519
+ ("ziad" nil (z iy1 ax d))
520
+ ("cmu" nil (s iy1 eh1 m y uw1))
521
+ ("hua" nil (w aa1))
522
+ ("kornel" nil (k ao r n eh1 l))
523
+ ("lavie" nil (l ax v iy1))
524
+ ("raux" nil (r ow1))
525
+ ("maeda" nil (m ay1 d ax))
526
+ ("portune" nil (p ao r t uw1 n))
527
+ ("stefanie" nil (s t eh1 f ae n iy))
528
+ ("woszczyna" nil (w aa zh iy1 n ax0))
529
+ ("waveform" nil (w ey1 v f ao1 r m))
530
+ ("waveforms" nil (w ey1 v f ao1 r m z))
531
+ ("dataset" nil (d ey1 t ah0 s eh1 t))
532
+ ("datasets" nil (d ey1 t ah0 s eh1 t s))
533
+
534
+ ;; Added July 2008
535
+ ("hermione" n (hh er m ay1 ax n iy))
536
+ ("tidied" v (t ay1 d ax d))
537
+ ("tidies" v (t ay1 d ax z))
538
+ ("untidy" v (ax n t ay1 d iy))
539
+ ("tidying" v (t ay1 d iy ih ng))
540
+ ("putting" v (p uh1 t ih ng))
541
+ ("iphone" n (ay1 f ow n))
542
+ ("ipod" n (ay1 p aa d))
543
+ ("untie" v (ax n t ay1))
544
+ ("kilohertz" n (k ih1 l ax hh er1 t s))
545
+ ("linux" n (l ih1 n ah k s))
546
+ ("miso" n (m iy1 s ow))
547
+ ("gaijin" n (g ay1 jh ih n))
548
+ ("uk" n (y uw1 k ey1))
549
+ ("katie" n (k ey1 t iy))
550
+ ("waistcoat" n (w ey1 s t k ow t))
551
+ ("tarpaulin" n (t aa r p ao1 l ih n))
552
+ ("trestle" n (t r eh1 s ax l))
553
+ ("trestles" n (t r eh1 s ax l z))
554
+ ("clunits" n (s iy1 eh1 l y uw1 n ih t s))
555
+ ("clustergen" n (k l ah1 s t er jh eh1 n))
556
+ ("libravox" n (l iy1 b r ax v aa1 k s))
557
+ ("peeped" n (p iy1 p t))
558
+ ("goodbyes" n (g uh1 d b ay1 z))
559
+ ("spearmint" j (s p iy1 r m ih0 n t))
560
+ ("surfed" v (s er1 f t))
561
+ ("unpacked" v (ax0 n p ae1 k t))
562
+ ("cookstove" v (k uh1 k s t ow1 v))
563
+ ("woodstove" v (w uh1 d s t ow1 v))
564
+ ("watertight" j (w ao1 t er0 t ay0 t))
565
+ ("youtube" n (y uw1 t uw1 b))
566
+ ("schlep" v (sh l eh1 p))
567
+ ("drawknife" n (d r ao1 n ay0 f))
568
+ ("wifi" n (w ay1 f ay1))
569
+ ("waterproof" j (w ao1 t er0 p r uw0 f))
570
+ ("understair" j (ah1 n d er0 s t ey1 r))
571
+ ("wikipedia" n (w ih0 k ih0 p iy1 d iy0 ax0))
572
+ ("smelt" v (s m eh1 l t))
573
+
574
+
CosyVoice-ttsfrd/resource/festival/dicts/cmu/cmulex.scm ADDED
@@ -0,0 +1,466 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1998 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Definition of Festival lexicon, derived from CMUDICT-0.4
35
+ ;;;
36
+
37
+ (defvar cmulexdir (path-append lexdir "cmu"))
38
+
39
+ (require 'pos)
40
+
41
+ (set! cmulex_version "2.0 (from 0.4) July 2008")
42
+
43
+ (define (cmu_lts_function word feats)
44
+ "(cmu_lts_function word feats)
45
+ Function called for CMULEX when word is not found in lexicon. Uses
46
+ LTS rules trained from the original lexicon, and lexical stress
47
+ prediction rules."
48
+ (require 'lts)
49
+ (if (not (boundp 'cmu_lts_rules))
50
+ (load (path-append cmulexdir "cmu_lts_rules.scm")))
51
+ (let ((dcword (downcase word))
52
+ (syls) (phones))
53
+ (if (string-matches dcword "[a-z]*")
54
+ (begin
55
+ (set! phones (lts_predict dcword cmu_lts_rules))
56
+ (set! syls (cmulex_mosyl_phstress phones))
57
+ )
58
+ (set! syls nil))
59
+ (list word nil syls)))
60
+
61
+ (define (cmulex_addenda)
62
+ "(cmulex_addenda)
63
+ Add entries to the current lexicon (radio/darpa). These are basically
64
+ words that are not in the CMU lexicon."
65
+ ; (lex.add.entry '("a's" n (((ey z) 1))))
66
+ ; (lex.add.entry '("e's" n (((iy z) 1))))
67
+ ; (lex.add.entry '("i's" n (((ay z) 1))))
68
+ ; (lex.add.entry '("o's" n (((ow z) 1))))
69
+ ; (lex.add.entry '("u's" n (((y uw z) 1))))
70
+ ; (lex.add.entry '("r's" n (((aa r z) 1))))
71
+ (lex.add.entry '("worf" n (((w ao r f) 1))))
72
+ (lex.add.entry '("t" n (((t iy) 1))))
73
+ (lex.add.entry '("I'll" v (((ay l) 1))))
74
+ (lex.add.entry '("it's" v (((ih t s) 1))))
75
+ (lex.add.entry '("don't" v (((d ow n t) 1))))
76
+ (lex.add.entry '("didn't" v (((d ih) 1) ((d ah n t) 0))))
77
+ (lex.add.entry '("isn't" v (((ih) 1) ((z ah n t) 0))))
78
+ (lex.add.entry '("doesn't" v (((d ah) 1) ((z ah n t) 0))))
79
+ (lex.add.entry '("hasn't" v (((hh ae) 1) ((z ah n t) 0))))
80
+ (lex.add.entry '("that's" v (((dh ae t s) 1))))
81
+ (lex.add.entry '("won't" v (((w ow n t) 1))))
82
+ (lex.add.entry '("aren't" v (((aa r n t) 1))))
83
+ (lex.add.entry '("there's" v (((dh er z) 1))))
84
+ (lex.add.entry '("we're" v (((w iy r) 1))))
85
+ (lex.add.entry '("wouldn't" v (((w uh) 1) ((d ah n t) 0))))
86
+ (lex.add.entry '("wasn't" v (((w aa) 1) ((z ah n t) 0))))
87
+ (lex.add.entry '("they're" v (((dh er) 1))))
88
+ (lex.add.entry '("weren't" v (((w er n t) 1))))
89
+ (lex.add.entry '("i'm" v (((ay m) 1))))
90
+ (lex.add.entry '("he's" v (((hh iy z) 1))))
91
+ (lex.add.entry '("you're" v (((y uw r) 1))))
92
+ (lex.add.entry '("haven't" v (((hh ae) 1) ((v ah n t) 0))))
93
+ (lex.add.entry '("we've" v (((w iy v) 1))))
94
+ (lex.add.entry '("i've" v (((ay v) 1))))
95
+ (lex.add.entry '("hadn't" v (((hh ae) 1) ((d ah n t) 0))))
96
+ (lex.add.entry '("they've" v (((dh ey v) 1))))
97
+ (lex.add.entry '("shouldn't" v (((sh uh) 1) ((d ah n t) 0))))
98
+ (lex.add.entry '("I'd" v (((ay d) 1))))
99
+ (lex.add.entry '("they'll" v (((dh ey l) 1))))
100
+ (lex.add.entry '("you've" v (((y uw v) 1))))
101
+ (lex.add.entry '("you'll" v (((y uw l) 1))))
102
+ (lex.add.entry '("I'll" v (((ay l) 1))))
103
+ (lex.add.entry '("we'd" v (((w iy d) 1))))
104
+ (lex.add.entry '("he'd" v (((hh iy d) 1))))
105
+ (lex.add.entry '("he'll" v (((hh iy l) 1))))
106
+ (lex.add.entry '("they'd" v (((dh ey d) 1))))
107
+ (lex.add.entry '("you'd" v (((y uw d) 1))))
108
+ (lex.add.entry '("it'll" v (((ih t) 1) ((ah l) 0))))
109
+ (lex.add.entry '("who've" v (((hh uw v) 1))))
110
+ (lex.add.entry '("ain't" v (((ey n t) 1))))
111
+ (lex.add.entry '("needn't" v (((n iy) 1) ((d ah n t) 0))))
112
+ (lex.add.entry '("she'd" v (((sh iy d) 1))))
113
+ (lex.add.entry '("who'd" v (((hh uw d) 1))))
114
+ (lex.add.entry '("she'll" v (((sh iy l) 1))))
115
+ (lex.add.entry '("there'll" v (((dh er l) 1))))
116
+ (lex.add.entry '("there'd" v (((dh er d) 1))))
117
+ (lex.add.entry '("it'd" v (((ih t) 1) ((ah d) 0))))
118
+ (lex.add.entry '("who'll" v (((hh uw l) 1))))
119
+ (lex.add.entry '("that'll" v (((dh ae t l) 1))))
120
+ (lex.add.entry '("mightn't" v (((m ay) 1) ((t ah n t) 0))))
121
+ (lex.add.entry '("would've" v (((w uh d) 1) ((ah v) 0))))
122
+ (lex.add.entry '("mustn't" v (((m ah) 1) ((s ah n t) 0))))
123
+ (lex.add.entry '("how'd" v (((hh ow d) 1))))
124
+ (lex.add.entry '("could've" v (((k uh d) 1) ((ah v) 0))))
125
+
126
+ (lex.add.entry '("couldn't" v (((k uh) 1) ((d ah n t) 0))))
127
+ (lex.add.entry '("can't" v (((k ae n t) 1))))
128
+ (lex.add.entry '("we'll" v (((w iy l) 1))))
129
+
130
+ (lex.add.entry '("uk" n (((y uw) 1) ((k ey) 1))))
131
+ (lex.add.entry '("w" n (((d ah b) 1) ((ah l) 0) ((y uw) 1))))
132
+ (lex.add.entry '("bought" v (((b ao t) 1))))
133
+ (lex.add.entry '("edinburgh" n (((eh d) 1) ((ah n) 0) ((b ax ) 0) ((r ow) 0))))
134
+ (lex.add.entry '("non-intoxicating" () (((n aa n) 1) ((ih n t) 0) ((aa k) 1) ((s ih k) 0) ((ey t) 1) ((ih ng) 0))))
135
+ (lex.add.entry '("AT&T" n (((ey) 1) ((t iy) 1) ((ah n d) 0) ((t iy) 1))))
136
+ (lex.add.entry
137
+ '("cepstra" n (((k eh p) 1) ((s t r aa) 0))))
138
+ (lex.add.entry
139
+ '("cepstral" n (((k eh p) 1) ((s t r ah l) 0))))
140
+ (lex.add.entry
141
+ '("cepstrum" n (((k eh p) 1) ((s t r ah m) 0))))
142
+ (lex.add.entry
143
+ '("diphone" n (((d ay) 1) ((f ow n) 0))))
144
+ (lex.add.entry
145
+ '("diphones" n (((d ay) 1) ((f ow n s) 0))))
146
+ (lex.add.entry
147
+ '("emacs" n (((iy) 1) ((m ae k s) 0))))
148
+ (lex.add.entry
149
+ '("hz" n (((hh eh r t z) 1))))
150
+ (lex.add.entry
151
+ '("khz" n (((k ih) 1) ((l ax) 0) ((hh eh r t z) 1))))
152
+ (lex.add.entry
153
+ '("lived" v (((l ih v d) 1))))
154
+ (lex.add.entry
155
+ '("Ltd" n (((l ih m) 1) ((ah t) 0) ((ah d) 0))))
156
+ (lex.add.entry
157
+ '("Mrs" n (((m ih s) 1) ((ah s) 0))))
158
+ (lex.add.entry
159
+ '("mhz" n (((m eh) 1) ((g ax) 0) ((hh eh r t z) 1))))
160
+ (lex.add.entry
161
+ '("NASA" n (((n ae) 1) ((s ax) 1))))
162
+ (lex.add.entry
163
+ '("lead" led (((l eh d) 1))))
164
+ (lex.add.entry
165
+ '("lead" liid (((l iy d) 1))))
166
+ (lex.add.entry
167
+ '("pasadena" n (((p ae s) 1) ((ae d) 0) ((iy n) 1) ((ax) 0))))
168
+ (lex.add.entry
169
+ '("read" red (((r eh d) 1))))
170
+ (lex.add.entry
171
+ '("read" riid (((r iy d) 1))))
172
+ (lex.add.entry
173
+ '("reuters" n (((r oy t) 1) ((er z) 0))))
174
+ (lex.add.entry
175
+ '("and" aend (((ae n d) 1))))
176
+ (lex.add.entry
177
+ '("and" ahnd (((ax n d) 0))))
178
+ (lex.add.entry
179
+ '("for" faor (((f ao r) 1))))
180
+ (lex.add.entry
181
+ '("for" fer (((f er) 0))))
182
+ (lex.add.entry
183
+ '("to" to (((t ax) 0))))
184
+ (lex.add.entry
185
+ '("to" tuu (((t uw) 1))))
186
+ ; (lex.add.entry
187
+ ; '("usually" r (((y uw zh) 1) ((ax l) 0) ((iy) 0))))
188
+ ; (lex.add.entry
189
+ ; '("usual" r (((y uw zh) 1) ((ax l) 0))))
190
+ ; (lex.add.entry
191
+ ; '("Jan" n (((jh ae n y) 1) ((uw) 0) ((eh r) 1) ((iy) 0))))
192
+ ; (lex.add.entry
193
+ ; '("Feb" n (((f eh b) 1) ((r ax) 0) ((er) 0) ((iy) 0))))
194
+ (lex.add.entry
195
+ '("February" n (((f eh b) 1) ((r ax) 0) ((er) 0) ((iy) 0))))
196
+ ; (lex.add.entry
197
+ ; '("Mar" n (((m aa r ch) 0))))
198
+ ; (lex.add.entry
199
+ ; '("Apr" n (((ey p) 1) ((r ah l) 0))))
200
+ ; (lex.add.entry
201
+ ; '("Jun" n (((jh uw n) 1))))
202
+ ; (lex.add.entry
203
+ ; '("Jul" n (((jh uw l) 1) ((ay) 1))))
204
+ ; (lex.add.entry
205
+ ; '("Aug" n (((aa g) 1) ((ah s t) 0))))
206
+ ; (lex.add.entry
207
+ ; '("Sep" n (((s eh p) 0) ((t eh m b) 1) ((er) 0))))
208
+ ; (lex.add.entry
209
+ ; '("Sept" n (((s eh p) 0) ((t eh m b) 1) ((er) 0))))
210
+ ; (lex.add.entry
211
+ ; '("Oct" n (((aa k) 0) ((t ow b) 1) ((er) 0))))
212
+ ; (lex.add.entry
213
+ ; '("Nov" n (((n ow v) 0) ((eh m b) 1) ((er) 0))))
214
+ ; (lex.add.entry
215
+ ; '("Dec" n (((d ih s) 0) ((eh m b) 1) ((er) 0))))
216
+ (lex.add.entry
217
+ '("'s" pos (((ax z) 0))))
218
+ (lex.add.entry
219
+ '("'s" n (((ax z) 0))))
220
+ (lex.add.entry
221
+ '("*" n (((ae s) 1) ((t er) 0) ((ih s k) 0)) ((pos "K6%"))))
222
+ (lex.add.entry
223
+ '("%" n (((p er) 1) ((s eh n t) 1)) ((pos "K9%"))))
224
+ (lex.add.entry
225
+ '("&" n (((ae n d) 0))))
226
+ (lex.add.entry
227
+ '("$" n (((d aa l) 1) ((er) 0))))
228
+ (lex.add.entry
229
+ '("#" n (((hh ae sh) 1))))
230
+ (lex.add.entry
231
+ '("@" n (((ae t) 1))))
232
+ (lex.add.entry
233
+ '("+" n (((p l ah s) 0)) ((pos "K7%" "OA%" "T-%"))))
234
+ (lex.add.entry
235
+ '("^" n (((k eh r) 1) ((eh t) 0)) ((pos "K6$"))))
236
+ (lex.add.entry
237
+ '("~" n (((t ih l d) 1) ((ah) 0)) ((pos "K6$"))))
238
+ (lex.add.entry
239
+ '("=" n (((iy k) 1) ((w ah l) 0))))
240
+ (lex.add.entry
241
+ '("/" n (((s l ae sh) 1))))
242
+ (lex.add.entry
243
+ '("\\" n (((b ae k) 1) ((s l ae sh) 1))))
244
+ (lex.add.entry
245
+ '("_" n (((ah n d) 1) ((er s) 0) ((k ao r) 1))))
246
+ (lex.add.entry
247
+ '("|" n (((v er t) 1) ((ih k) 0) ((ah l) 0) ((b aa r) 1))))
248
+ (lex.add.entry
249
+ '(">" n ((( g r ey ) 1) ((t er) 0) ((dh ae n) 1))))
250
+ (lex.add.entry
251
+ '("<" n ((( l eh s) 1) ((dh ae n) 1))))
252
+ (lex.add.entry
253
+ '("[" n ((( l eh f t) 1) ((b r ae k) 1) ((ih t) 0))))
254
+ (lex.add.entry
255
+ '("]" n ((( r ay t) 1) ((b r ae k) 1) ((ih t) 0))))
256
+ (lex.add.entry
257
+ '(" " n (((s p ey s) 1))))
258
+ (lex.add.entry
259
+ '("\t" n (((t ae b) 1))))
260
+ (lex.add.entry
261
+ '("\n" n (((n uw) 1) ((l ay n) 1))))
262
+ (lex.add.entry '("." punc nil))
263
+ (lex.add.entry '("'" punc nil))
264
+ (lex.add.entry '(":" punc nil))
265
+ (lex.add.entry '(";" punc nil))
266
+ (lex.add.entry '("," punc nil))
267
+ (lex.add.entry '("-" punc nil))
268
+ (lex.add.entry '("\"" punc nil))
269
+ (lex.add.entry '("`" punc nil))
270
+ (lex.add.entry '("?" punc nil))
271
+ (lex.add.entry '("!" punc nil))
272
+ (lex.add.entry
273
+ '(before ()
274
+ (((b iy) 0) ((f ao r) 1))))
275
+ )
276
+
277
+ (define (cmulex_map_sylstructure syls)
278
+ (unwind-protect
279
+ (begin
280
+ (mapcar
281
+ (lambda (s1 s2)
282
+ (list s1 (cadr s2)))
283
+ (cmulex_syllabify_maxonset
284
+ (apply
285
+ append
286
+ (mapcar car syls)))
287
+ syls))
288
+ (begin
289
+ (format t "Failed to resyllabify %l\n" syls)
290
+ syls))
291
+ )
292
+
293
+ (define (cmulex_mosyl_phstress phones)
294
+ (set! xxx (mapcar
295
+ (lambda (syl)
296
+ (set! stress 0)
297
+ (list
298
+ (mapcar
299
+ (lambda (p)
300
+ (cond
301
+ ((string-matches p "[aeiou@].*1")
302
+ (set! stress 1)
303
+ (intern (substring p 0 (- (length p) 1))))
304
+ ((string-matches p "[aeiou@].*0")
305
+ (set! stress 0)
306
+ (intern (substring p 0 (- (length p) 1))))
307
+ (t
308
+ (intern p))))
309
+ (mapcar string-append syl))
310
+ stress))
311
+ (cmulex_syllabify_maxonset phones)))
312
+ ; (format t "%l\n%l\n" phones xxx)
313
+ xxx
314
+
315
+ )
316
+
317
+ (define (cmulex_syllabify_maxonset phones)
318
+ (cmulex_syllabify_maxonset2
319
+ (mapcar intern (reverse phones))
320
+ nil nil)
321
+ )
322
+
323
+ (define (cmulex_syllabify_maxonset2 phones syl syls)
324
+ "(cmulex_syllabify_maxonset phones)
325
+ Syllabify by maximum onset. phones is given in reverse order"
326
+ ; (format t "csm2 phones %l syl %l syls %l\n"
327
+ ; phones syl syls)
328
+ (cond
329
+ ((null phones)
330
+ (if syl
331
+ (cons (reverse syl) syls)
332
+ syls))
333
+ ((null (cmulex_has_vowel phones)) ;; safety case
334
+ ;; could some weird onset we've never seen before
335
+ (cons
336
+ (append (reverse phones) syl)
337
+ syls))
338
+ ((null (string-matches (car phones) "[aeiou@].*")) ;; a vowel
339
+ (cmulex_syllabify_maxonset2
340
+ (cdr phones)
341
+ (cons (car phones) syl)
342
+ syls))
343
+ (t ;; is a vowel
344
+ (let ((onset (cmulex_maxonset (cdr phones))))
345
+ (cmulex_syllabify_maxonset2
346
+ (nth_cdr (+ 1 (length onset)) phones)
347
+ nil
348
+ (cons (append onset (list (car phones)) syl) syls))))))
349
+
350
+ (define (cmulex_has_vowel p)
351
+ (cond
352
+ ((null p) nil)
353
+ ((string-matches (car p) "[aeiou@].*") ;; a vowel
354
+ t)
355
+ (t
356
+ (cmulex_has_vowel (cdr p)))))
357
+
358
+ (define (cmulex_maxonset phones)
359
+ (cond
360
+ ((string-matches (car phones) "[aeiou@].*")
361
+ nil)
362
+ ((string-equal (car phones) "ng") ;; only non-syl-initial phone
363
+ nil)
364
+ ((null phones) nil)
365
+ ((and (> (length phones) 2)
366
+ (member (list (car (cddr phones))
367
+ (cadr phones)
368
+ (car phones)
369
+ )
370
+ cmulex_tri_onsets))
371
+ (list (car (cddr phones))
372
+ (cadr phones)
373
+ (car phones)
374
+ ))
375
+ ((and (> (length phones) 1)
376
+ (member (list (cadr phones)
377
+ (car phones))
378
+ cmulex_di_onsets))
379
+ (list (cadr phones)
380
+ (car phones)))
381
+ (t
382
+ (list (car phones)))))
383
+
384
+ (set! cmulex_tri_onsets
385
+ '(
386
+ (s t r)
387
+ (s p y)
388
+ (s p r)
389
+ (s p l)
390
+ (s k y)
391
+ (s k w)
392
+ (s k r)
393
+ (s k l)
394
+ ))
395
+
396
+ (set! cmulex_di_onsets
397
+ '(
398
+ (z w)
399
+ (z l)
400
+ (v y)
401
+ (v r)
402
+ (v l)
403
+ (th w)
404
+ (th r)
405
+ (t y)
406
+ (t w)
407
+ ; (t s)
408
+ (t r)
409
+ (sh w)
410
+ (sh r)
411
+ (sh n)
412
+ (sh m)
413
+ (sh l)
414
+ (s w)
415
+ (s v)
416
+ (s t)
417
+ (s r)
418
+ (s p)
419
+ (s n)
420
+ (s m)
421
+ (s l)
422
+ (s k)
423
+ (s f)
424
+ (p y)
425
+ (p w)
426
+ (p r)
427
+ (p l)
428
+ (n y)
429
+ (m y)
430
+ (m r)
431
+ (l y)
432
+ (k y)
433
+ (k w)
434
+ (k r)
435
+ (k l)
436
+ (hh y)
437
+ (hh w)
438
+ (hh r)
439
+ (hh l)
440
+ (g y)
441
+ (g w)
442
+ (g r)
443
+ (g l)
444
+ (f y)
445
+ (f r)
446
+ (f l)
447
+ (d y)
448
+ (d w)
449
+ (d r)
450
+ (b y)
451
+ (b w)
452
+ (b r)
453
+ (b l)
454
+ ))
455
+
456
+ (set! lex_syllabification (list cmulex_mosyl_phstress))
457
+
458
+ (lex.create "cmu")
459
+ (lex.set.compile.file (path-append cmulexdir "cmudict-0.4.out"))
460
+ (lex.set.phoneset "radio")
461
+ (lex.set.lts.method 'cmu_lts_function)
462
+ (lex.set.pos.map english_pos_map_wp39_to_wp20)
463
+ (cmulex_addenda)
464
+
465
+ (provide 'cmulex)
466
+
CosyVoice-ttsfrd/resource/festival/dicts/oald/COPYING ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ Centre for Speech Technology Research
3
+ University of Edinburgh, UK
4
+ Copyright (c) 1996,1997
5
+ All Rights Reserved.
6
+
7
+ Permission to use, copy, modify, distribute this software and its
8
+ documentation for research, educational and individual use only, is
9
+ hereby granted without fee, subject to the following conditions:
10
+ 1. The code must retain the above copyright notice, this list of
11
+ conditions and the following disclaimer.
12
+ 2. Any modifications must be clearly marked as such.
13
+ 3. Original authors' names are not deleted.
14
+ This software may not be used for commercial purposes without
15
+ specific prior written permission from the authors.
16
+
17
+ THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK
18
+ DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING
19
+ ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT
20
+ SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE
21
+ FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
22
+ WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
23
+ AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
24
+ ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
25
+ THIS SOFTWARE.
26
+ Centre for Speech Technology Research
27
+ University of Edinburgh, UK
28
+ Copyright (c) 1996,1997
29
+ All Rights Reserved.
30
+
31
+ Permission is hereby granted, free of charge, to use and distribute
32
+ this software and its documentation without restriction, including
33
+ without limitation the rights to use, copy, modify, merge, publish,
34
+ distribute, sublicense, and/or sell copies of this work, and to
35
+ permit persons to whom this work is furnished to do so, subject to
CosyVoice-ttsfrd/resource/festival/dicts/oald/ali.oald.lex ADDED
The diff for this file is too large to render. See raw diff
 
CosyVoice-ttsfrd/resource/festival/dicts/oald/oald_extensions.scm ADDED
@@ -0,0 +1,1704 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1998 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Extra items that have sufficient frequency and are pronounced
35
+ ;;; wrongly that are to be added to the compiled version of
36
+ ;;; the OALD lexicon
37
+ ;;;
38
+ ;;; This list we derived from various text databases and actually
39
+ ;;; listening
40
+
41
+ ( "Jones" n (jh ou1 n z) )
42
+ ( "Merrill" n (m e1 r i l) )
43
+ ( "Guber" n (g u1 b @) )
44
+ ( "Drexel" n (d r e1 k s @ l) )
45
+ ( "de" in (d @) )
46
+ ( "Noriega" n (n oo r i ei1 g @) )
47
+ ( "Lehman" n (l ei1 m @ n) )
48
+ ( "Salomon" n (s a1 l @ m @ n) )
49
+ ( "pretax" n (p r ii1 t a1 k s) )
50
+ ( "Reagan" n (r ei1 g @ n) )
51
+ ( "Hugo" n (h y uu1 g o) )
52
+ ( "Nasdaq" n (n a1 s d e k) )
53
+ ( "Telerate" n (t e1 l i r ei t) )
54
+ ( "PaineWebber" n (p ei1 n - w e1 b @) )
55
+ ( "totaled" n (t ou1 t @ l d) )
56
+ ( "Qintex" n (k w i1 n t e k s) )
57
+ ( "Conn" n (k o1 n) )
58
+ ( "Asia" n (ei1 sh i@) )
59
+ ( "Saatchi" n (s a1 t ch ii) )
60
+ ( "Mitsubishi" n (m i1 t s uu b ii1 sh ii) )
61
+ ( "Mae" n (m ei1) )
62
+ ( "Paribas" n (p a1 r @ b a s) )
63
+ ( "Intel" n (i1 n t e l) )
64
+ ( "McCaw" n (m @ k oo1) )
65
+ ( "Citicorp" n (s i1 t ii - k oo1 p) )
66
+ ( "automotive" n (oo1 t ou m ou1 t i v) )
67
+ ( "Compaq" n (k o1 m p a1 k) )
68
+ ( "Beijing" n (b ei1 zh i1 ng) )
69
+ ( "rumors" n (r uu1 m @@ z) )
70
+ ( "Exxon" n (e1 k s o n) )
71
+ ( "Phelan" n (f ii1 l a n) )
72
+ ( "Chrysler" n (k r ai1 s l @) )
73
+ ( "Oakland" n (ou1 k l @ n d) )
74
+ ( "insurers" n (i n sh u@1 r @ z) )
75
+ ( "Thomson" n (t o1 m s @ n) )
76
+ ( "Stearns" n (s t @@1 n z) )
77
+ ( "Coors" n (k u@1 z) )
78
+ ( "USAir" n (y uu1 e1 s e@1) )
79
+ ( "Rockefeller" n (r o1 k @ f e l @) )
80
+ ( "Ogilvy" n (ou1 g i l v ii) )
81
+ ( "pre" n (p r ii1) )
82
+ ( "Bache" n (b a1 k) )
83
+ ( "Mitchell" n (m i1 t ch @ l) )
84
+ ( "Davis" n (d ei1 v i s) )
85
+ ( "Nekoosa" n (n e k uu1 s @) )
86
+ ( "Nikkei" n (n ii1 k k ei1) )
87
+ ( "McDonald" n (m @ k d o1 n @ l d) )
88
+ ( "Marshall" n (m aa1 sh @ l) )
89
+ ( "aide" n (ei1 d) )
90
+ ( "CenTrust" n (s e1 n - t r uh s t) )
91
+ ( "BellSouth" n (b e1 l s au1 th) )
92
+ ( "Thompson" n (t o1 m s @ n) )
93
+ ( "Donaldson" n (d o1 n @ l d s @ n) )
94
+ ( "totaling" n (t ou1 t @ l i ng) )
95
+ ( "insurer" n (i n sh u@1 r @) )
96
+ ( "Campbell" n (k a1 m b @ l) )
97
+ ( "favorable" k (f ei1 v r @ b @ l) )
98
+ ( "canceled" j (k a1 n s @ l d) )
99
+ ( "Du" in (d uu) )
100
+ ( "Fujitsu" n (f uu jh i1 t s uu) )
101
+ ( "Yeargin" n (y e@1 g i n) )
102
+ ( "Daiwa" n (d ai1 w a) )
103
+ ( "Bancorp" n (b a1 n k oo p) )
104
+ ( "Viacom" n (v ai1 @ k o m) )
105
+ ( "Vegas" n (v ei1 g @ s) )
106
+ ( "Steinhardt" n (s t ai1 n h aa1 t) )
107
+ ( "Rica" n (r ii1 k @) )
108
+ ( "Icahn" n (ii1 k a n) )
109
+ ( "Ferranti" n (f @ r a1 n t ii) )
110
+ ( "Ortega" n (oo r t ei1 g @) )
111
+ ( "Merieux" n (m e1 r y uu) )
112
+ ( "Spielvogel" n (s p ii1 l v ou1 k @ l) )
113
+ ( "Gonzalez" n (g o n z a1 l @ s) )
114
+ ( "Engelken" n (e1 ng g @ l k e n) )
115
+ ( "Delmed" n (d e1 l m e1 d) )
116
+ ( "Corry" n (k o1 r ii) )
117
+ ( "longtime" n (l o1 ng t ai1 m) )
118
+ ( "Saab" n (s a1 b) )
119
+ ( "Rey" n (r ei1) )
120
+ ( "Nabisco" n (n @ b i1 s k ou) )
121
+ ( "Disney" n (d i1 z n ii) )
122
+ ( "repurchase" n (r ii1 p @@1 ch i s) )
123
+ ( "mainframes" n (m ei1 n f r ei1 m z) )
124
+ ( "Reynolds" n (r e1 n @ l d z) )
125
+ ( "Nomura" n (n ou m uu1 r @) )
126
+ ( "Galileo" n (g a1 l @ l ei1 ou) )
127
+ ( "Cohen" n (k ou1 e n) )
128
+ ( "Allianz" n (@ l ai1 a n z) )
129
+ ( "biotechnology" n (b ai1 ou t e1 k n o1 l @ jh ii) )
130
+ ( "Jenrette" n (zh o1 n r e1 t) )
131
+ ( "Bethlehem" n (b e1 th l i h e m) )
132
+ ( "turnaround" n (t @@1 n @ r au1 n d) )
133
+ ( "rumored" n (r uu1 m @@ d) )
134
+ ( "cleanup" n (k l ii1 n uh1 p) )
135
+ ( "Wedtech" n (w e1 d t e1 k) )
136
+ ( "Texaco" n (t e1 k s @ k ou1) )
137
+ ( "Sullivan" n (s uh1 l @ v @ n) )
138
+ ( "Roberts" n (r o1 b @ t s) )
139
+ ( "Merksamer" n (m @@1 k s @ m @) )
140
+ ( "Hahn" n (h a1 n) )
141
+ ( "Giuliani" n (jh uu1 l ii a1 n ii) )
142
+ ( "theater" n (th i@1 t @) )
143
+ ( "Unisys" n (y uu1 n i s i1 s) )
144
+ ( "Monieson" n (m o1 n @ s @ n) )
145
+ ( "Diego" n (d i ei1 g ou) )
146
+ ( "Quotron" n (k w o1 t r o n) )
147
+ ( "Maynard" n (m ei1 n aa d) )
148
+ ( "Marcos" n (m aa1 k o s) )
149
+ ( "Jose" n (h ou1 z ei) )
150
+ ( "Boren" n (b oo1 r e n) )
151
+ ( "signaled" n (s i1 g n @ l d) )
152
+ ( "lackluster" n (l a1 k - l uh1 s t @) )
153
+ ( "Stein" n (s t ii1 n) )
154
+ ( "Bartlett" n (b aa1 t l e t) )
155
+ ( "Amex" n (a1 m e1 k s) )
156
+ ( "aides" n (ei1 d z) )
157
+ ( "Trelleborg" n (t r e1 l @ b @@ g) )
158
+ ( "Saudi" n (s au1 d ii) )
159
+ ( "Olivetti" n (o1 l i v e1 t ii) )
160
+ ( "Microsoft" n (m ai1 k r ou s o1 f t) )
161
+ ( "McGraw" n (m @ g r oo1) )
162
+ ( "Deutsche" n (d oi1 t ch @) )
163
+ ( "Courter" n (k oo1 t @) )
164
+ ( "Conner" n (k o1 n @) )
165
+ ( "Cathay" n (k a1 th ei1) )
166
+ ( "polyethylene" n (p o1 l ii e1 th i l ii n) )
167
+ ( "Quebecor" n (k w i1 b e k @) )
168
+ ( "Puerto" n (p w e@1 t ou) )
169
+ ( "Mikhail" n (m ii1 k ai l) )
170
+ ( "Colo" n (k o1 l ou) )
171
+ ( "refinancing" n (r ii f ai1 n @ n s i ng) )
172
+ ( "Societe" n (s @1 s ai1 @ t ei) )
173
+ ( "Robertson" n (r o1 b @ t s s @ n) )
174
+ ( "Pfeiffer" n (f ai1 f @) )
175
+ ( "Oppenheimer" n (o1 p @ n h ai m @) )
176
+ ( "Evans" n (e1 v @ n z) )
177
+ ( "Dataproducts" n (d ei1 t @ - p r o1 d uh k t s) )
178
+ ( "Bradstreet" n (b r a1 d - s t r ii1 t) )
179
+ ( "Antonio" n (a n t ou1 n y ou) )
180
+ ( "franchisees" n (f r a1 n ch ai z ii1 z) )
181
+ ( "ex" n (e1 k s) )
182
+ ( "arbitragers" n (aa1 b @ t r ei1 jh @@ z) )
183
+ ( "Motorola" n (m ou t @ r ou1 l @) )
184
+ ( "Macmillan" n (m @ k m i1 l @ n) )
185
+ ( "Jacobson" n (jh ei1 k @ b s @ n) )
186
+ ( "Indosuez" n (i1 n d ou s uu1 e z) )
187
+ ( "Hughes" n (h y uu1 z) )
188
+ ( "Florio" n (f l oo r i ou) )
189
+ ( "Spiegel" n (s p ii1 g @ l) )
190
+ ( "Rico" n (r ii1 k ou) )
191
+ ( "Palo" n (p a1 l ou) )
192
+ ( "Madison" n (m a1 d @ s @ n) )
193
+ ( "HealthVest" n (h e1 l th - v e1 s t) )
194
+ ( "Fuji" n (f uu1 jh ii) )
195
+ ( "Deloitte" n (d e1 l oi t) )
196
+ ( "Amoco" n (a1 m ou k ou) )
197
+ ( "equaling" n (ii1 k w @ l i ng) )
198
+ ( "Wallach" n (w oo1 l @ k) )
199
+ ( "Upjohn" n (uh1 p jh o n) )
200
+ ( "Tele" n (t e1 l ii) )
201
+ ( "Rubicam" n (r uu1 b @ k a1 m) )
202
+ ( "Pemex" n (p e1 m e1 k s) )
203
+ ( "Meredith" n (m e1 r @ d @ th) )
204
+ ( "Lorenzo" n (l oo r e1 n z ou) )
205
+ ( "Levine" n (l @ v ai1 n) )
206
+ ( "Ingersoll" n (i1 ng g @ s @ l) )
207
+ ( "Eurocom" n (y uu1 r @ k o m) )
208
+ ( "Chevrolet" n (sh e1 v r @ l ei) )
209
+ ( "federally" n (f e1 d r @ l ii) )
210
+ ( "anymore" n (e1 n ii m oo1 r) )
211
+ ( "Toshiba" n (t ou sh ii1 b @) )
212
+ ( "Suisse" n (s w ii1 s) )
213
+ ( "Sansui" n (s a1 n s uu ii1) )
214
+ ( "Panzhihua" n (p a1 n z ii1 w a1) )
215
+ ( "Goldberg" n (g ou1 l d b @@ g) )
216
+ ( "Garcia" n (g aa s ii1 @) )
217
+ ( "Eli" n (ii1 l ai) )
218
+ ( "Della" n (d e1 l @) )
219
+ ( "Buick" n (b y uu1 i k) )
220
+ ( "debacle" n (d ei b uh1 k @ l) )
221
+ ( "cholesterol" n (k o l e1 s t r @ l) )
222
+ ( "Vitro" n (v ii1 t r ou) )
223
+ ( "Revco" n (r e1 v k ou) )
224
+ ( "Nashua" n (n a1 sh w @) )
225
+ ( "Mitsui" n (m i t s uu1 ii) )
226
+ ( "Lexus" n (l e1 k s @ s) )
227
+ ( "Healthcare" n (h e1 l th - k e@ r) )
228
+ ( "Ginnie" n (jh i1 n ii) )
229
+ ( "Carlos" n (k aa1 l @ s) )
230
+ ( "Benson" n (b e1 n s @ n) )
231
+ ( "Avery" n (ei1 v r ii) )
232
+ ( "Achenbaum" n (a1 k @ n b oo m) )
233
+ ( "Nynex" n (n ai1 n e k s) )
234
+ ( "Murray" n (m uh1 r ii) )
235
+ ( "Morishita" n (m o1 r ii1 sh t @) )
236
+ ( "Minpeco" n (m i1 n p e k ou) )
237
+ ( "McDonough" n (m @ k d o1 n @) )
238
+ ( "Geigy" n (g ei1 jh ii) )
239
+ ( "Drabinsky" n (d r a b i1 n s k ii) )
240
+ ( "Dingell" n (d i1 ng @ l) )
241
+ ( "petrochemical" n (p e1 t r @ k e1 m @ k @ l) )
242
+ ( "nonetheless" n (n uh1 n dh @ l e1 s) )
243
+ ( "catalog" n (k a1 t @ l o g) )
244
+ ( "Yamaichi" n (y a1 m @ ii1 ch ii) )
245
+ ( "Sugarman" n (sh uu1 g @ m @ n) )
246
+ ( "Steinberg" n (s t ai1 n b @@ g) )
247
+ ( "Shevardnadze" n (sh e1 v @ d n a1 d z ii) )
248
+ ( "Russell" n (r uh1 s @ l) )
249
+ ( "Roh" n (r ou1) )
250
+ ( "McDonnell" n (m @ k d o1 n @ l) )
251
+ ( "Lyonnais" n (l ei1 @ n ei z) )
252
+ ( "Kravis" n (k r a1 v @ s) )
253
+ ( "Gould" n (g uu1 l d) )
254
+ ( "Edelman" n (ei1 d @ l m @ n) )
255
+ ( "Dreyfus" n (d r ei1 f @ s) )
256
+ ( "Coniston" n (k o1 n @ s t @ n) )
257
+ ( "Cineplex" n (s i1 n @ p l e k s) )
258
+ ( "Antar" n (a1 n t aa) )
259
+ ( "Annualized" n (a1 n y uu l ai z d) )
260
+ ( "salespeople" n (s ei1 l s p ii p @ l) )
261
+ ( "pharmaceuticals" n (f aa1 m @ s y uu1 t @ k @ l z) )
262
+ ( "layoffs" n (l ei1 o1 f s) )
263
+ ( "issuance" n (i sh uu1 a n s) )
264
+ ( "intraday" n (i1 n t r @ d ei1) )
265
+ ( "exercisable" n (e1 k s @ r s ai1 z @ b @ l) )
266
+ ( "Weisfield" n (w ii1 s f ii1 l d) )
267
+ ( "Unocal" n (y uu1 n @ k a l) )
268
+ ( "Straszheim" n (s t r a1 z h ai m) )
269
+ ( "Microsystems" n (m ai1 k r ou s i1 s t @ m z) )
270
+ ( "Hambrecht" n (h a1 m b r e k t) )
271
+ ( "Generale" n (jh e n e r a1 l) )
272
+ ( "Friedman" n (f r ii1 d m @ n) )
273
+ ( "Fresenius" n (f r ii s e1 n i uh s) )
274
+ ( "Contel" n (k o1 n t e l) )
275
+ ( "Bennett" n (b e1 n i t) )
276
+ ( "Azoff" n (a1 z o f) )
277
+ ( "Arabia" n (@ r ei1 b i@) )
278
+ ( "rubles" n (r uu1 b @ l z) )
279
+ ( "routinely" n (r uu t ii1 n l ii) )
280
+ ( "Unilever" n (y uu1 n @ l ii1 v @ r) )
281
+ ( "Telesis" n (t e1 l ii s i s) )
282
+ ( "Skase" n (s k a1 s ii) )
283
+ ( "Manuel" n (m a1 n w e l) )
284
+ ( "Mancuso" n (m a n k y uu1 s ou) )
285
+ ( "Hearst" n (h @@1 s t) )
286
+ ( "DeConcini" n (d @ k o n ch ii1 n ii) )
287
+ ( "Chiron" n (ch ai1 r o n) )
288
+ ( "Caribbean" n (k a1 r @ b ii1 @ n) )
289
+ ( "Bogart" n (b ou1 g aa t) )
290
+ ( "rumor" n (r uu1 m @ r) )
291
+ ( "noncallable" n (n o1 n k oo1 l @ b @ l) )
292
+ ( "industrywide" n (i1 n d uh s t r ii w ai1 d) )
293
+ ( "Weiss" n (w ai1 s) )
294
+ ( "Tucson" n (t uu1 s o n) )
295
+ ( "Rosenthal" n (r ou1 s @ n t a l) )
296
+ ( "Rogers" n (r o1 jh @@ z) )
297
+ ( "Roderick" n (r o1 d @ r i k) )
298
+ ( "Rican" n (r ii1 k a n) )
299
+ ( "Prebon" n (p r ii1 b o n) )
300
+ ( "Owen" n (ou1 e n) )
301
+ ( "Mercedes" n (m @@ s ei1 d ii s) )
302
+ ( "Intelogic" n (i1 n t e l l o1 jh i k) )
303
+ ( "HomeFed" n (h ou1 m f e1 d) )
304
+ ( "Fossett" n (f o1 s i t) )
305
+ ( "Cocom" n (k ou1 k o1 m) )
306
+ ( "Cheney" n (ch ei1 n ii) )
307
+ ( "Capel" n (k a p e1 l) )
308
+ ( "Cabrera" n (k @ b r e1 r @) )
309
+ ( "Blumenfeld" n (b l uu1 m @ n f e l d) )
310
+ ( "pressured" n (p r e1 sh @ r d) )
311
+ ( "pachinko" n (p @ ch i1 ng k ou) )
312
+ ( "conferees" n (k o1 n f @ r ii1 z) )
313
+ ( "audio" n (oo1 d i ou) )
314
+ ( "Weyerhaeuser" n (w ai1 r h au1 s @ r) )
315
+ ( "Schwarz" n (sh w oo1 t z) )
316
+ ( "Schering" n (sh e1 r i ng) )
317
+ ( "Sandinista" n (s a1 n @ n ii1 s t @) )
318
+ ( "Renault" n (r e1 n ou) )
319
+ ( "Petrie" n (p ii1 t r ii) )
320
+ ( "Mosbacher" n (m o1 s b a k @ r) )
321
+ ( "Mario" n (m a1 r i ou) )
322
+ ( "Kleinwort" n (k l ai1 n w oo t) )
323
+ ( "Hyundai" n (h uu1 n d ai) )
324
+ ( "Genentech" n (jh e1 n @ n t e1 k) )
325
+ ( "Busch" n (b uu1 sh) )
326
+ ( "Bofors" n (b ou1 f @ @ s) )
327
+ ( "Barre" n (b aa1 r) )
328
+ ( "ruble" n (r uu1 b @ l) )
329
+ ( "nonprofit" n (n o1 n - p r o1 f i t) )
330
+ ( "buildup" n (b i1 l d uh1 p) )
331
+ ( "breakup" n (b r ei1 k uh1 p) )
332
+ ( "ampersand" n (a1 m p @ s a n d) )
333
+ ( "TransCanada" n (t r a1 n s - k a1 n @ d @1) )
334
+ ( "Superfund" n (s uu1 p @ - f uh n d) )
335
+ ( "Sandinistas" n (s a1 n @ n ii1 s t @ s) )
336
+ ( "Posner" n (p o1 z n @ r) )
337
+ ( "Papandreou" n (p a1 p @ n d r ii ou) )
338
+ ( "Odeon" n (ou1 d i @ n) )
339
+ ( "Nathan" n (n ei1 th @ n) )
340
+ ( "McGovern" n (m @ g uh1 v @@ n) )
341
+ ( "Matra" n (m a1 t r @) )
342
+ ( "Lorin" n (l oo1 r @ n) )
343
+ ( "Klein" n (k l ai1 n) )
344
+ ( "Goodyear" n (g u1 d y e@1) )
345
+ ( "Eaton" n (ii1 t @ n) )
346
+ ( "Donoghue" n (d o1 n @ h y uu) )
347
+ ( "DaPuzzo" n (d @ p uu1 z ou) )
348
+ ( "Barclays" n (b aa1 k l i z) )
349
+ ( "semiannual" n (s e1 m i - a1 n y u@ l) )
350
+ ( "maneuver" n (m a n uu1 v @ r) )
351
+ ( "gridlock" n (g r i1 d l o1 k) )
352
+ ( "defenses" n (d @ f e1 n s @ s) )
353
+ ( "anytime" n (e1 n ii t ai m) )
354
+ ( "annualized" n (a1 n y u@ l ei z d) )
355
+ ( "Trinova" n (t r i n ou1 v @) )
356
+ ( "Suzuki" n (s u z uu1 k ii) )
357
+ ( "Shapiro" n (sh @ p ii1 r ou) )
358
+ ( "Schwartz" n (sh w oo1 t z) )
359
+ ( "Savaiko" n (s a v ai1 k ou) )
360
+ ( "Salinas" n (s a l ii1 n @ s) )
361
+ ( "Rio" n (r ii1 ou) )
362
+ ( "Refcorp" n (r e1 f k oo p) )
363
+ ( "Mehta" n (m ei1 t @) )
364
+ ( "McMoRan" n (m @ k m oo1 r @ n) )
365
+ ( "Matsushita" n (m a t s uu1 sh t @) )
366
+ ( "Luzon" n (l uu1 z o n) )
367
+ ( "Lufthansa" n (l uu f t h a1 n s @) )
368
+ ( "Interprovincial" n (i1 n t @@ - p r o v i1 n sh @ l) )
369
+ ( "Hoechst" n (h ou1 sh t) )
370
+ ( "Giovanni" n (jh i ou v a1 n ii) )
371
+ ( "Erbamont" n (e1 r b @ m o n t) )
372
+ ( "Emhart" n (e1 m h aa t) )
373
+ ( "Daewoo" n (d ei1 w uu) )
374
+ ( "Corr" n (k o1 r) )
375
+ ( "Chivas" n (ch ii1 v @ s) )
376
+ ( "Burger" n (b @@1 g @) )
377
+ ( "Broadway" n (b r oo1 d w ei) )
378
+ ( "Alaskan" n (@ l a1 s k @ n) )
379
+ ( "palladium" n (p @ l ei1 d i uh m) )
380
+ ( "installment" n (i n s t oo1 l m @ n t) )
381
+ ( "franchisee" n (f r a1 n ch ai z ii1) )
382
+ ( "Wertheim" n (w e1 r th ai m) )
383
+ ( "Uniroyal" n (y uu1 n i r oi1 @ l) )
384
+ ( "Southam" n (s au1 th @ m) )
385
+ ( "Sinyard" n (s i1 n y aa d) )
386
+ ( "Quayle" n (k w ei1 l) )
387
+ ( "Pontiac" n (p o1 n t i a k) )
388
+ ( "Nuovo" n (n uu1 @ v ou) )
389
+ ( "MedChem" n (m e1 d k e m) )
390
+ ( "Mahfouz" n (m aa1 f uu z) )
391
+ ( "Lomas" n (l ou1 m a s) )
392
+ ( "Kangyo" n (k a1 ng g y ou) )
393
+ ( "Iverson" n (ai1 v @@ s @ n) )
394
+ ( "Graham" n (g r ei1 uh m) )
395
+ ( "Giorgio" n (g ii1 oo r g ii1 ou) )
396
+ ( "Gelbart" n (g e1 l b aa t) )
397
+ ( "Gabelli" n (g a b e1 l ii) )
398
+ ( "Ellis" n (e1 l i s) )
399
+ ( "Donuts" n (d ou1 n uh t s) )
400
+ ( "videocassette" n (v i1 d i ou - k a s e1 t) )
401
+ ( "petrochemicals" n (p e1 t r @ k e1 m @ k @ l z) )
402
+ ( "multinational" n (m uh1 l t i - n a1 sh @ n @ l) )
403
+ ( "minivans" n (m i1 n i v a n z) )
404
+ ( "gubernatorial" n (g uu1 b @@ n a t oo1 r i@ l) )
405
+ ( "franchisers" n (f r a1 n ch ai z @@ z) )
406
+ ( "advancers" n (@ d v a1 n s @@ s) )
407
+ ( "Zoete" n (z ou1 e t) )
408
+ ( "Weil" n (w ei1 l) )
409
+ ( "Torrijos" n (t o r r ii1 y o s) )
410
+ ( "Templeton" n (t e1 m p @ l t @ n) )
411
+ ( "Suez" n (s uu1 e z) )
412
+ ( "Sperry" n (s p e1 r ii) )
413
+ ( "Seita" n (s ei1 t @) )
414
+ ( "Scania" n (s k a1 n i @) )
415
+ ( "Salinger" n (s a1 l i n g @ r) )
416
+ ( "Rudolph" n (r uu1 d o l f) )
417
+ ( "Rostenkowski" n (r o1 s t @ n k au1 s k ii) )
418
+ ( "Pierre" n (p i1 e@) )
419
+ ( "Peugeot" n (p y uu1 zh ou) )
420
+ ( "Pearce" n (p i@1 s) )
421
+ ( "Ortiz" n (o r t ii1 z) )
422
+ ( "Ogden" n (o1 g d i n) )
423
+ ( "OBrien" n (ou b r ai1 @ n) )
424
+ ( "Neuberger" n (n y uu1 b @@ g @ r) )
425
+ ( "Murata" n (m uu r a1 t @) )
426
+ ( "McNamee" n (m a k n @ m ii1) )
427
+ ( "McLennan" n (m @ k l e1 n @ n) )
428
+ ( "Mayer" n (m ei1 @) )
429
+ ( "Mattel" n (m a t e1 l) )
430
+ ( "Macintosh" n (m a1 k i n t o sh) )
431
+ ( "MacDonald" n (m a k d o1 n @ l d) )
432
+ ( "Lumpur" n (l uu1 m p @) )
433
+ ( "Kuala" n (k w a1 l @) )
434
+ ( "Honolulu" n (h o1 n @ l uu1 l uu) )
435
+ ( "Harrison" n (h a1 r i s @ n) )
436
+ ( "Haas" n (h aa1 s) )
437
+ ( "Francois" n (f r o1 n s w @@ z) )
438
+ ( "Cupertino" n (k uu p @ t ii1 n ou) )
439
+ ( "CityFed" n (s i1 t ii f e1 d) )
440
+ ( "Capcom" n (k a1 p k o m) )
441
+ ( "Cananea" n (k @ n a1 n i@) )
442
+ ( "Caltrans" n (k a1 l t r a n s) )
443
+ ( "Beebes" n (b ii1 b s) )
444
+ ( "Bebear" n (b ii1 b e@ r) )
445
+ ( "Avis" n (ei1 v i s) )
446
+ ( "Aoun" n (a1 uu n) )
447
+ ( "Adm" n (a1 d m i r @ l) )
448
+ ( "Abramson" n (ei1 b r @ m s @ n) )
449
+ ( "theaters" n (th i@1 t @@ z) )
450
+ ( "resale" n (r ii s ei1 l) )
451
+ ( "pretrial" n (p r ii1 t r ai1 @ l) )
452
+ ( "midyear" n (m i1 d y @@ r) )
453
+ ( "en" in (o ng) )
454
+ ( "elite" n (i l ii1 t) )
455
+ ( "divisive" n (d @ v ai1 z i v) )
456
+ ( "bono" n (b ou1 n ou) )
457
+ ( "archrival" n (aa1 ch r ai v @ l) )
458
+ ( "acquirer" n (@ k w ai1 r @) )
459
+ ( "Vogelstein" n (v ou1 g @ l s t ii n) )
460
+ ( "Utsumi" n (uu t s uu1 m ii) )
461
+ ( "Touche" n (t uu sh ei1) )
462
+ ( "Tenneco" n (t e1 n @ k ou) )
463
+ ( "Sumitomo" n (s uu1 m ii t ou1 m ou) )
464
+ ( "Sohmer" n (s ou1 m @ r) )
465
+ ( "Shakespeare" n (sh ei1 k s p i@ r) )
466
+ ( "Samsung" n (s a1 m s uh n g) )
467
+ ( "Perlman" n (p @@1 l m @ n) )
468
+ ( "Patel" n (p @ t e1 l) )
469
+ ( "Oy" n (o1 i1) )
470
+ ( "Omaha" n (ou1 m @ h @) )
471
+ ( "ONeill" n (ou n ii1 l) )
472
+ ( "Mueller" n (m uu1 l @ r) )
473
+ ( "Mitterrand" n (m ii1 t @ r a n) )
474
+ ( "Mideast" n (m i1 d ii s t) )
475
+ ( "Michel" n (m ai1 k @ l) )
476
+ ( "Mehl" n (m e1 l) )
477
+ ( "Kume" n (k uu1 m e) )
478
+ ( "Gutfreund" n (g u1 t f r e n d) )
479
+ ( "Geographic" n (jh i1 ou g r a1 f i k) )
480
+ ( "Fujis" n (f uu1 jh ii z) )
481
+ ( "Freres" n (f r e1 r z) )
482
+ ( "Fazio" n (f a1 t s i ou) )
483
+ ( "Einhorn" n (ai1 n h oo n) )
484
+ ( "Egon" n (ii1 g o n) )
485
+ ( "Cruz" n (k r uu1 z) )
486
+ ( "Comex" n (k o1 m e k s) )
487
+ ( "Carnegie" n (k a r n ei1 g ii) )
488
+ ( "Braniff" n (b r a1 n i f) )
489
+ ( "Aquino" n (@ k ii1 n ou) )
490
+ ( "Aichi" n (ai ii1 ch ii) )
491
+ ( "Ahmanson" n (aa1 m @ n s @ n) )
492
+ ( "statewide" n (s t ei1 t w ai d) )
493
+ ( "sitcom" n (s i1 t k o m) )
494
+ ( "refinance" n (r ii f ai1 n a n s) )
495
+ ( "postwar" n (p ou s t w oo1 r) )
496
+ ( "playoffs" n (p l ei o1 f s) )
497
+ ( "outpaced" n (au t p ei1 s d) )
498
+ ( "midtown" n (m i d t au1 n) )
499
+ ( "maneuvering" n (m a n uu1 v @ r i ng) )
500
+ ( "makeup" n (m ei1 k uh1 p) )
501
+ ( "indexation" n (i1 n d e k s ei1 sh @ n) )
502
+ ( "hemorrhaging" n (h e1 m @ r @ jh i ng) )
503
+ ( "closings" n (k l ou1 s i ng z) )
504
+ ( "Xtra" n (e1 k s t r @) )
505
+ ( "Wachter" n (w o1 ch @ r) )
506
+ ( "Trudeau" n (t r uu1 d ou) )
507
+ ( "Tharp" n (th aa1 p) )
508
+ ( "Stephens" n (s t ii1 f @ n z) )
509
+ ( "Sino" n (s ai1 n ou) )
510
+ ( "Seymour" n (s ii1 m o r) )
511
+ ( "Reitman" n (r ai1 t m @ n) )
512
+ ( "Ramirez" n (r @ m ii1 r e z) )
513
+ ( "Petrolane" n (p e1 t r ou l ei n) )
514
+ ( "Omni" n (o1 m n ii) )
515
+ ( "Ohbayashi" n (ou1 b a y a1 sh ii) )
516
+ ( "Oakes" n (ou1 k s) )
517
+ ( "Nuys" n (n oi1 z) )
518
+ ( "Nguyen" n (n uu1 y e n) )
519
+ ( "Mateo" n (m @ t ei1 ou) )
520
+ ( "Maclean" n (m @ k l ii n) )
521
+ ( "Kia" n (k ii1 @) )
522
+ ( "Kerr" n (k e r) )
523
+ ( "Kawasaki" n (k a1 w @ s a1 k ii) )
524
+ ( "Kao" n (k a1 ou) )
525
+ ( "Kalikow" n (k a1 l @ k ou) )
526
+ ( "Jacques" n (zh a1 k) )
527
+ ( "Jacobs" n (jh ei1 k @ b s) )
528
+ ( "Institut" n (i1 n s t i t uu t) )
529
+ ( "Hawaiian" n (h @ w ai1 @ n) )
530
+ ( "Garratt" n (g a1 r i t) )
531
+ ( "Elliott" n (e1 l ii i t) )
532
+ ( "Ehrlich" n (e1 r l i k) )
533
+ ( "Eduard" n (e1 d w aa d) )
534
+ ( "Dai" n (d ai1) )
535
+ ( "Conant" n (k o1 n @ n t) )
536
+ ( "Bruno" n (b r uu1 n ou) )
537
+ ( "wildlife" n (w ai1 l d l ai f) )
538
+ ( "wastewater" n (w ei1 s t w oo t @ r) )
539
+ ( "unraveling" n (uh n r a1 v l i ng) )
540
+ ( "tritium" n (t r i1 t i uh m) )
541
+ ( "semiannually" n (s e1 m i a1 n y u@ l) )
542
+ ( "salmonella" n (s a1 l m @ n e1 l @) )
543
+ ( "reoffered" n (r ii1 o1 f @@ d) )
544
+ ( "reinvest" n (r ii1 i n v e1 s t) )
545
+ ( "merchandising" n (m @@1 ch @ n d ai1 z i ng) )
546
+ ( "favorably" n (f ei1 v r @ b l ii) )
547
+ ( "duo" n (d y uu1 ou) )
548
+ ( "dialysis" n (d ai a1 l @ s i s) )
549
+ ( "chlorofluorocarbons" n (k l o1 r @ f l o1 r @ k aa1 b @ n z) )
550
+ ( "Yamatake" n (y a1 m @ t a1 k e) )
551
+ ( "Wichita" n (w i1 ch @ t @) )
552
+ ( "Valdez" n (v a1 l d e z) )
553
+ ( "Unicorp" n (y uu1 n i k oo p) )
554
+ ( "Trustcorp" n (t r uh1 s t k oo p) )
555
+ ( "Toubro" n (t ou1 b r ou) )
556
+ ( "Tiananmen" n (t i@1 n @ n m e1 n) )
557
+ ( "Tesoro" n (t e s ou1 r ou) )
558
+ ( "Tae" n (t ei1) )
559
+ ( "Stalinist" n (s t a1 l @ n @ s t) )
560
+ ( "SsangYong" n (s a1 ng y o1 ng) )
561
+ ( "Shidler" n (sh ai1 d l @ r) )
562
+ ( "Schuster" n (sh u1 s t @ r) )
563
+ ( "Schlumberger" n (sh l uh1 m b @@ zh ei) )
564
+ ( "Schaeffer" n (sh ei1 f @ r) )
565
+ ( "Sagan" n (s ei1 g @ n) )
566
+ ( "Roberti" n (r @ b @@1 t ii) )
567
+ ( "Rahn" n (r aa1 n) )
568
+ ( "Rafale" n (r @ f a1 l ii) )
569
+ ( "Quina" n (k ii1 n @) )
570
+ ( "Parenthood" n (p e@1 r @ n t h uu d) )
571
+ ( "Novello" n (n o v e1 l ou) )
572
+ ( "Morgenzon" n (m o1 r g @ n z @ n) )
573
+ ( "Moines" n (m oi1 n) )
574
+ ( "Michelin" n (m i1 sh @ l a n) )
575
+ ( "Meyer" n (m ai1 @ r) )
576
+ ( "McGee" n (m @ g ii1) )
577
+ ( "Marous" n (m a1 r ou s) )
578
+ ( "Managua" n (m @ n a1 g w @) )
579
+ ( "Macrodantin" n (m a1 k r ou d a1 n t i n) )
580
+ ( "Knopf" n (k n o1 p f) )
581
+ ( "Jayark" n (jh ei1 y aa k) )
582
+ ( "Izvestia" n (i s v e1 s t i @) )
583
+ ( "Ichi" n (ii1 ch ii) )
584
+ ( "Hyman" n (h ai1 m @ n) )
585
+ ( "Hirsch" n (h @@1 sh) )
586
+ ( "Himont" n (h ai1 m @ n t) )
587
+ ( "Hercules" n (h @@1 k y uu l ii z) )
588
+ ( "Heinemann" n (h ai1 n m @ n) )
589
+ ( "Hathaway" n (h a1 th @ w ei) )
590
+ ( "Guterman" n (g u1 t @@ m @ n) )
591
+ ( "Gitano" n (g ii t a1 n ou) )
592
+ ( "Gibson" n (g i1 b s @ n) )
593
+ ( "Geduld" n (g e1 d uh l d) )
594
+ ( "Gatward" n (g a1 t w aa d) )
595
+ ( "Gartner" n (g aa1 t n @ r) )
596
+ ( "Fujisawa" n (f uu1 jh ii s a1 w @) )
597
+ ( "FirstSouth" n (f i1 r s t s au th) )
598
+ ( "Finkelstein" n (f i1 n k @ l s t ii n) )
599
+ ( "Fernandez" n (f @@ n a1 n d e z) )
600
+ ( "Faberge" n (f a1 b e r jh ei) )
601
+ ( "Detrex" n (d e1 t r e k s) )
602
+ ( "Carolinas" n (k a1 r @ l ai1 n @ s) )
603
+ ( "Candela" n (k @ n d e1 l @) )
604
+ ( "Calloway" n (k a1 l @ w ei) )
605
+ ( "Cablevision" n (k ei1 b @ v i1 zh @ n) )
606
+ ( "Cabernet" n (k a1 b @ n ei) )
607
+ ( "Byrd" n (b @@1 d) )
608
+ ( "Bundesbank" n (b uu1 n d @ s b a1 n k) )
609
+ ( "Bumiputra" n (b u1 m i p u1 t r @) )
610
+ ( "Boyer" n (b oi1 @ r) )
611
+ ( "Boise" n (b oi1 s ii) )
612
+ ( "Albuquerque" n (a1 l b @ k @@1 k ii) )
613
+ ( "Accor" n (@ k o1 r) )
614
+ ( "Acadia" n (@ k ei1 d i @) )
615
+ ( "unfocused" n (uh n f ou1 k @ s d) )
616
+ ( "superpower" n (s uu1 p @ p au1 @ r) )
617
+ ( "subcompact" n (s uh1 b k o1 m p a k t) )
618
+ ( "spreadsheet" n (s p r e1 d sh ii t) )
619
+ ( "signaling" n (s i1 g n @ l i ng) )
620
+ ( "setup" n (s e t uh1 p) )
621
+ ( "repurchased" n (r ii1 p @@1 ch i s d) )
622
+ ( "refinanced" n (r ii f ai1 n a n s d) )
623
+ ( "pressuring" n (p r e1 sh @ r i ng) )
624
+ ( "paychecks" n (p ei1 ch e k s) )
625
+ ( "math" n (m a1 th) )
626
+ ( "favorites" n (f ei1 v r i t s) )
627
+ ( "destabilizing" n (d ii s t ei1 b @ l ai1 z i ng) )
628
+ ( "crocidolite" n (k r ou s i1 d @ l ai t) )
629
+ ( "counterrevolutionary" n (k au1 n t @@ - r e1 v @ l uu1 sh @ n r ii) )
630
+ ( "colorful" n (k uh1 l @ r f @ l) )
631
+ ( "bureaus" n (b y uu1 r ou s) )
632
+ ( "broadcasters" n (b r oo1 d k a s t @@ s) )
633
+ ( "Yasuda" n (y a1 s u d @) )
634
+ ( "Yamamoto" n (y a1 m a m ou1 t ou) )
635
+ ( "Volkswagen" n (v o1 l k s w a g @ n) )
636
+ ( "Veraldi" n (v e r a1 l d ii) )
637
+ ( "Unice" n (y uu1 n i s) )
638
+ ( "Tulsa" n (t uh1 l s @) )
639
+ ( "Travis" n (t r a1 v @ s) )
640
+ ( "Traviata" n (t r a v ii a1 t @) )
641
+ ( "Threlkeld" n (th r e1 l k e l d) )
642
+ ( "Stronach" n (s t r o1 n @ k) )
643
+ ( "Stovall" n (s t o1 v a l) )
644
+ ( "Staten" n (s t a1 t @ n) )
645
+ ( "Soweto" n (s u w e1 t ou) )
646
+ ( "Siegel" n (s ii1 g @ l) )
647
+ ( "Shattuck" n (sh a1 t uh k) )
648
+ ( "Segundo" n (s e g u1 n d ou) )
649
+ ( "SciMed" n (s ai1 m e1 d) )
650
+ ( "Schulof" n (sh u1 l o f) )
651
+ ( "Schneider" n (sh n ai1 d @ r) )
652
+ ( "Schaefer" n (sh ai1 f @ r) )
653
+ ( "Salem" n (s ei1 l @ m) )
654
+ ( "Safeco" n (s ei1 f k ou) )
655
+ ( "Ruvolo" n (r uu1 v ou l ou) )
656
+ ( "Rianta" n (r ii a1 n t @) )
657
+ ( "Prix" n (p r ii1) )
658
+ ( "Petipa" n (p e t ii1 p @) )
659
+ ( "Paxus" n (p a1 k s @ s) )
660
+ ( "Paulo" n (p au1 l ou) )
661
+ ( "Parsow" n (p aa1 s ou) )
662
+ ( "Ovcharenko" n (o1 v ch a r e1 n k ou) )
663
+ ( "Orlando" n (o r l a1 n d d ou) )
664
+ ( "Novell" n (n ou v e1 l) )
665
+ ( "Nissho" n (n ii1 sh ou) )
666
+ ( "Nikon" n (n ii1 k o n) )
667
+ ( "Ngoc" n (n o1 k) )
668
+ ( "Movieline" n (m uu1 v ii l ai n) )
669
+ ( "Montagu" n (m o1 n t @ g y uu) )
670
+ ( "Mitsukoshi" n (m i1 t s uu k ou1 sh ii) )
671
+ ( "Milpitas" n (m i l p ii1 t @ s) )
672
+ ( "McGill" n (m @ g i1 l) )
673
+ ( "McDermott" n (m @ k d @@1 m i t) )
674
+ ( "Maxtor" n (m a1 k s t o r) )
675
+ ( "Marietta" n (m a1 r i e1 t @) )
676
+ ( "Maier" n (m ai1 @ r) )
677
+ ( "Maguire" n (m @ g w ai1 @ r) )
678
+ ( "Machiguenga" n (m a1 ch i g w e1 n g @) )
679
+ ( "Lowell" n (l au1 @ l) )
680
+ ( "Lorillard" n (l o1 r @ l aa d) )
681
+ ( "Lonrho" n (l oo1 n r ou) )
682
+ ( "Lentjes" n (l e1 n t y i z) )
683
+ ( "Kozinski" n (k o z i1 n s k ii) )
684
+ ( "Kloves" n (k l ou1 v z) )
685
+ ( "Jerusalem" n (jh e r uu1 s @ l e m) )
686
+ ( "Iwai" n (ii w ai1) )
687
+ ( "Immunex" n (i1 m y uu n e k s) )
688
+ ( "Ikegai" n (ii1 k e g ai) )
689
+ ( "Homestake" n (h ou1 m s t ei k) )
690
+ ( "Heinz" n (h ai1 n z) )
691
+ ( "Harlem" n (h aa1 l @ m) )
692
+ ( "Hallwood" n (h oo1 l w u d) )
693
+ ( "Haagen" n (h aa1 g @ n) )
694
+ ( "Geva" n (jh e1 v @) )
695
+ ( "Georgetown" n (jh oo1 jh t au n) )
696
+ ( "GenCorp" n (jh e1 n k oo p) )
697
+ ( "Garzarelli" n (g aa1 z @ r e1 l ii) )
698
+ ( "Fremont" n (f r ii1 m o n t) )
699
+ ( "Forster" n (f oo1 s t @ r) )
700
+ ( "Fleischmann" n (f l e1 sh m @ n) )
701
+ ( "Fischer" n (f i1 sh @ r) )
702
+ ( "Fidel" n (f i d e1 l) )
703
+ ( "Feshbach" n (f e1 sh b a k) )
704
+ ( "Ferruzzi" n (f e r uu1 t s ii) )
705
+ ( "Felipe" n (f e l ii1 p) )
706
+ ( "Deseret" n (d e1 s e r i t) )
707
+ ( "Davidson" n (d ei1 v i d s @ n) )
708
+ ( "Cooperman" n (k uu1 p @@ m a n) )
709
+ ( "Chez" n (sh e1 z) )
710
+ ( "Canelo" n (k @ n e1 l ou) )
711
+ ( "Canaan" n (k ei1 n @ n) )
712
+ ( "Byrne" n (b @@ n) )
713
+ ( "Brody" n (b r ou1 d ii) )
714
+ ( "BroadBeach" n (b r oo1 d - b ii ch) )
715
+ ( "BioSciences" n (b ai1 ou - s ai1 @ n s @ s) )
716
+ ( "Barr" n (b aa1) )
717
+ ( "BancOklahoma" n (b a n k - ou1 k l @ h ou1 m @) )
718
+ ( "Asahi" n (a1 s a h ii) )
719
+ ( "Armenian" n (aa1 m i n i@ n) )
720
+ ( "Antarctica" n (@ n t aa1 k t i k @) )
721
+ ( "Anacomp" n (a1 n @ k o m p) )
722
+ ( "Ames" n (ei1 m z) )
723
+ ( "Amdahl" n (a1 m d a l) )
724
+ ( "Allday" n (oo1 l d ei1) )
725
+ ( "Ailes" n (ai1 l z) )
726
+ ( "whove" n (w ou v) )
727
+ ( "thirtysomething" n (th @@1 t ii - s uh1 m th i ng) )
728
+ ( "telemarketing" n (t e1 l ii - m aa1 k i t i ng) )
729
+ ( "telemarketers" n (t e1 l ii - m aa1 k i t @ r z) )
730
+ ( "sustainable" n (s @ s t ei1 n @ b @ l) )
731
+ ( "spiraling" n (s p ai1 r @ l i ng) )
732
+ ( "shoreline" n (sh oo1 r - l ai n) )
733
+ ( "retirees" n (r i t ai1 r ii1 z) )
734
+ ( "reseller" n (r ii1 s e l @ r) )
735
+ ( "reignited" n (r ii i g n ai1 t @ d) )
736
+ ( "refocusing" n (r ii f ou1 k @ s i ng) )
737
+ ( "refocus" n (r ii f ou1 k @ s) )
738
+ ( "pico" n (p ii1 k ou) )
739
+ ( "overreacting" n (ou1 v @@ - r ii a1 k t i ng) )
740
+ ( "onetime" n (w uh1 n - t ai1 m) )
741
+ ( "muni" n (m y uu1 n ii) )
742
+ ( "multinationals" n (m uh1 l t i - n a1 sh @ n @ l z) )
743
+ ( "maneuvers" n (m a n uu1 v @ r z) )
744
+ ( "haired" n (h e@1 d) )
745
+ ( "franchiser" n (f r a1 n ch ai z @ r) )
746
+ ( "franchised" n (f r a1 n ch ai z d) )
747
+ ( "feisty" n (f ai1 s t ii) )
748
+ ( "equaled" n (ii1 k w @ l d) )
749
+ ( "criminality" n (k r i1 m @ n a1 l i t ii) )
750
+ ( "coupe" n (k uu1 p ei) )
751
+ ( "confrontational" n (k o1 n f r @ n t ei1 sh @ n @ l) )
752
+ ( "catheter" n (k a1 th @ t @ r) )
753
+ ( "buyout" n (b ai1 au t) )
754
+ ( "armored" n (aa1 m @ r d) )
755
+ ( "aerobics" n (ei r ou1 b i k s) )
756
+ ( "acquirers" n (@ k w ai1 r @ r z) )
757
+ ( "academia" n (a1 k @ d ii1 m i @) )
758
+
759
+ ;;; no prizes for guessing where this came from
760
+
761
+ ( "Holmes" n (h ou1 m z) )
762
+ ( "Lestrade" n (l @ s t r aa1 d) )
763
+ ( "McGinty" n (m @ g i1 n t ii) )
764
+ ( "Stapleton" n (s t ei1 p @ l t @ n) )
765
+ ( "Baskerville" n (b a1 s k @ v i l) )
766
+ ( "Drebber" n (d r e1 b @ r) )
767
+ ( "Moriarty" n (m o1 r i aa1 t ii) )
768
+ ( "Conan" n (k o1 n @ n) )
769
+ ( "Tregennis" n (t r i g e1 n @ s) )
770
+ ( "Rucastle" n (r uu1 k a s @ l) )
771
+ ( "leading" n (l ii1 d i ng) )
772
+ ( "Ettie" n ( e1 t ii ))
773
+ ( "Carruthers" n ( k aa r uh1 th @@ z ) )
774
+ ( "McFarlane" n (m @ k f aa1 l i n) )
775
+ ( "Oldacre" n (ou1 l d - ei k @ r) )
776
+ ( "Alec" n (a1 l i k) )
777
+ ( "Lucas" n (l uu1 k @ s) )
778
+ ( "Cadogan" n (k @ d ou1 g @ n) )
779
+ ( "Huxtable" n (h uh1 k s t @ b @ l) )
780
+ ( "Athelney" n (a1 th @ l n ei) )
781
+ ( "daresay" n (d e@1 s ei) )
782
+ ( "Woolwich" n (w u1 l i ch) )
783
+ ( "Devonshire" n (d e1 v @ n sh @ r) )
784
+ ( "Trevelyan" n (t r @ v e1 l @ n) )
785
+ ( "Oberstein" n (ou1 b @ s t ai n) )
786
+ ( "Napoleon" n (n @ p ou1 l i @ n) )
787
+ ( "Dartmoor" n (d aa1 t m u@ r) )
788
+ ( "Sterndale" n (s t @@1 n d ei l) )
789
+ ( "Forrester" n (f o1 r @ s t @ r) )
790
+ ( "Woking" n (w ou1 k i ng) )
791
+ ( "Stapletons" n (s t ei1 p @ l t @ n s) )
792
+ ( "Mapleton" n (m ei1 p @ l t @ n) )
793
+ ( "Merryweather" n (m e1 r i w e dh @ r) )
794
+ ( "Gorgiano" n (jh oo jh i a1 n ou) )
795
+ ( "Beddoes" n (b e1 d ou s) )
796
+ ( "Prendergast" n (p r e1 n d @ g a s t) )
797
+ ( "Paddington" n (p a1 d i ng t @ n) )
798
+ ( "Thames" n (t e1 m z) )
799
+ ( "Charpentier" n (sh a r p e1 n t i ei) )
800
+ ( "halloa" n (h a l ou1 @) )
801
+ ( "Eustace" n (y uu1 s t @ s) )
802
+ ( "Baskervilles" n (b a1 s k @ v i l z) )
803
+ ( "merest" n (m ii1 r @ s t) )
804
+ ( "Mordecai" n (m oo1 d @ k ai) )
805
+ ( "Coburg" n (k ou1 b @@ g) )
806
+ ( "Swandam" n (s w o1 n d @ m) )
807
+ ( "Saxe" n (s a k s) )
808
+ ( "Grimesby" n (g r i1 m s b ii) )
809
+ ( "Esher" n (e1 sh @ r) )
810
+ ( "Eduardo" n (e d w aa1 d ou) )
811
+ ( "wagonette" n (w a1 g @ n e1 t) )
812
+ ( "Trelawney" n (t r @ l oo1 n ii) )
813
+ ( "Sydenham" n (s i1 d @ n h @ m) )
814
+ ( "Montpellier" n (m o m p e1 l i ei) )
815
+ ( "Lysander" n (l ai s a1 n d @ r) )
816
+ ( "Fairbairn" n (f e@1 b e@ n) )
817
+ ( "Altamont" n (a1 l t @ m o n t) )
818
+ ( "wagons" n (w a g @ n z) )
819
+ ( "outre" n (ou1 t r @) )
820
+ ( "moustached" n (m @ s t a1 sh d) )
821
+ ( "humouredly" n (h y uu1 m @ d l ii) )
822
+ ( "Silas" n (s ai1 l @ s) )
823
+ ( "Reuben" n (r uu1 b @ n) )
824
+ ( "Reichenbach" n (r i1 k @ n b a k) )
825
+ ( "Pietro" n (p ii e1 t r ou) )
826
+ ( "Heidegger" n (h ai1 d i g @ r) )
827
+ ( "Euston" n (y uu1 s t @ n) )
828
+ ( "Diogenes" n (d ai oo1 jh @ n ii z) )
829
+ ( "wagon" n (w a1 g @ n) )
830
+ ( "reentered" n (r ii e1 n t @ d) )
831
+ ( "alkaloid" n (a1 l k @ l oi d) )
832
+ ( "Trichinopoly" n (t r ai1 k i n o1 p @ l ii) )
833
+ ( "Streatham" n (s t r e1 t @ m) )
834
+ ( "Sophy" n (s ou1 f ii) )
835
+ ( "Meiringen" n (m e@1 r i ng @ n) )
836
+ ( "Greuze" n (g r uu1 z) )
837
+ ( "Cornelius" n (k o r n ii1 l i @ s) )
838
+ ( "fiancee" n (f i o1 n s ii1) )
839
+ ( "dyou" n (jh ou1) )
840
+ ( "bladed" n (b l ei1 d @ d) )
841
+ ( "aneurism" n (a1 n y uu r @ z m) )
842
+ ( "Vibart" n (v ai1 b aa t) )
843
+
844
+ ;;; Another set
845
+
846
+ ( "Mandela" n (m a n d e1 l @) )
847
+ ( "Johnson" n (jh o n s @ n) )
848
+ ( "Buchanan" n (b u k a1 n @ n) )
849
+ ( "Ceausescu" n (ch au ch e1 s k uu) )
850
+ ( "Schwarzkopf" n (sh w oo1 t s k o p f) )
851
+ ( "Aziz" n (@ z ii1 z) )
852
+ ( "Tyson" n (t ai1 s @ n) )
853
+ ( "Stalin" n (s t a1 l i n) )
854
+ ( "Cuomo" n (k uu1 m ou) )
855
+ ( "Rumania" n (r uu m ei1 n i @) )
856
+ ( "Powell" n (p au1 w @ l) )
857
+ ( "Viet" n (v i e1 t) )
858
+ ( "Kerrey" n (k e1 r ii) )
859
+ ( "Sununu" n (s @ n uu1 n uu) )
860
+ ( "Dukakis" n (d @ k a1 k @ s) )
861
+ ( "Lenin" n (l e1 n i n) )
862
+ ( "Chua" n (ch w a1) )
863
+ ( "Azerbaijan" n (a1 z @ b @ jh a1 n) )
864
+ ( "Riyadh" n (r ii1 a d) )
865
+ ( "Jesse" n (jh e1 s ii) )
866
+ ( "Hitler" n (h i1 t l @ r) )
867
+ ( "Tehran" n (t ei r a1 n) )
868
+ ( "Abu" n (a1 b uu) )
869
+ ( "Urquhart" n (@@1 k @@ t) )
870
+ ( "Mozart" n (m ou t s aa t) )
871
+ ( "Lieut" n (l e1 f t e1 n @ n t) )
872
+ ( "Faludi" n (f a l uu1 d ii) )
873
+ ( "Azerbaijanis" n (a1 z @@ b @ jh a1 n ii z) )
874
+ ( "Mubarak" n (m uu1 b @ r a k) )
875
+ ( "Le" dt (l @) )
876
+ ( "Botha" n (b ou1 t @) )
877
+ ( "Ukraine" n (y uu k r ei1 n) )
878
+ ( "antiwar" n (a1 n t i w oo1 r) )
879
+ ( "Likud" n (l i k uu1 d) )
880
+ ( "Duarte" n (d u aa1 t @) )
881
+ ( "Armenia" n (aa m ii1 n i @) )
882
+ ( "semiautomatic" n (s e1 m i oo1 t @ m a t i k) )
883
+ ( "Andrea" n (aa1 n d r i @) )
884
+ ( "Roosevelt" n (r ou z @ v e l t) )
885
+ ( "Luther" n (l u1 th @ r) )
886
+ ( "Armenians" n (aa m ii1 n i @ n z) )
887
+ ( "Bucharest" n (b u1 k @ r e s t) )
888
+ ( "Hirohito" n (h i1 r ou h ii1 t ou) )
889
+ ( "Modrow" n (m u1 d r ou) )
890
+ ( "Ayatullah" n (ai1 @ t uh1 l @) )
891
+ ( "Moynihan" n (m oi1 n @ h a n) )
892
+ ( "Gingrich" n (g i1 n g r i ch) )
893
+ ( "Securitate" n (s @ k y uu1 @ t ei t) )
894
+ ( "Mao" n (m au1) )
895
+ ( "Perez" n (p e1 r e z) )
896
+ ( "Kissinger" n (k i s @ n jh @ r) )
897
+ ( "Yankelovich" n (y a1 n k @ l o1 v i ch) )
898
+ ( "Steinem" n (s t ei n @ m) )
899
+ ( "Rashid" n (r a sh ii1 d) )
900
+ ( "Eisenhower" n (ai1 s @ n h au @ r) )
901
+ ( "Landsbergis" n (l a1 n d s b @@ g @ s) )
902
+ ( "Khrushchev" n (k r uu1 sh ch @ v) )
903
+ ( "Rodriguez" n (r o d r ii1 g e z) )
904
+ ( "Croatia" n (k r ou ei1 sh i@) )
905
+ ( "antiaircraft" n (a1 n t i e@1 k r a f t) )
906
+ ( "Rumanian" n (r uu m ei1 n i @ n) )
907
+ ( "Azerbaijani" n (a1 z @@ b @ jh a1 n ii) )
908
+ ( "Perot" n (p e r ou1) )
909
+ ( "FDIC" n (e1 f d i k) )
910
+ ( "Wagner" n (v aa1 g n @ r) )
911
+ ( "Amman" n (@ m a1 n) )
912
+ ( "mujahedin" n (m u jh a1 h @ d ii n) )
913
+ ( "genome" n (jh ii1 n ou m) )
914
+ ( "Nicolae" n (n i1 k ou l ai) )
915
+ ( "Mohammed" n (m @ h a1 m e d) )
916
+ ( "HIV" n (ei1 ch ai1 v ii1) )
917
+ ( "Aviv" n (@ v ii1 v) )
918
+ ( "quo" n (k w ou1) )
919
+ ( "intifadeh" n (i1 n t @ f a1 d @) )
920
+ ( "Monet" n (m o n ei1) )
921
+ ( "Karabakh" n (k a1 r @ b a k) )
922
+ ( "Hazelwood" n (h ei1 z @ l w u d) )
923
+ ( "Barrett" n (b a1 r i t) )
924
+ ( "furor" n ( f u r o1 r) )
925
+ ( "Tariq" n (t a1 r ii k) )
926
+ ( "Kazakhstan" n (k a1 z @ k s t a n) )
927
+ ( "Goodgame" n (g u1 d g ei m) )
928
+ ( "Pynchon" n (p i1 n ch @ n) )
929
+ ( "Ashe" n (a1 sh) )
930
+ ( "neo" n (n ii1 ou) )
931
+ ( "Rabta" n (r a1 b t @) )
932
+ ( "Pulitzer" n (p uu1 l i t z @ r) )
933
+ ( "Brezhnev" n (b r e1 zh n @ v) )
934
+ ( "Bhutto" n (b uu1 t ou) )
935
+ ( "Sidey" n (s ai1 d ii) )
936
+ ( "superpowers" n (s uu1 p @ p au @ r z) )
937
+ ( "Vaclav" n (v a1 z l a v) )
938
+ ( "Thom" n (t o1 m) )
939
+ ( "NIH" n (e1 n ai1 ei1 ch) )
940
+ ( "Isabelle" n (i1 s @ b e l) )
941
+ ( "Hosni" n (h o1 s n ii) )
942
+ ( "Slovenia" n (s l @ v ii1 n i @) )
943
+ ( "Milosevic" n (m i l o1 s @ v i ch) )
944
+ ( "MacLeod" n (m @ k l au1 d) )
945
+ ( "Friedrich" n (f r ii1 d r i k) )
946
+ ( "Timisoara" n (t ii1 m @ sh w a1 r @) )
947
+ ( "Tadzhikistan" n (t a d z i1 k i s t a n) )
948
+ ( "Salvadoran" n (s a1 l v @ d ou1 r @ n) )
949
+ ( "Rumanians" n (r uu m ei1 n i @ n z) )
950
+ ( "Perrier" n (p e1 r i ei) )
951
+ ( "Lautrec" n (l ou t r e1 k) )
952
+ ( "Kaifu" n (k ai1 f uu) )
953
+ ( "Heidi" n (h ai1 d ii) )
954
+ ( "Fahd" n (f aa d) )
955
+ ( "Disneyland" n (d i1 z n ii l a n d) )
956
+ ( "Courbet" n (k oo1 b ei) )
957
+ ( "Beckett" n (b e1 k i t) )
958
+ ( "Takeshita" n (t @ k e1 sh t @) )
959
+ ( "Khafji" n (k a1 f jh ii) )
960
+ ( "Daley" n (d ei1 l ii) )
961
+ ( "Clarence" n (k l a1 r @ n s) )
962
+ ( "firepower" n (f ai1 @ r p au1 @ r) )
963
+ ( "antiapartheid" n (a1 n t i @ p aa1 t h ai t) )
964
+ ( "Tobias" n (t ou b ai1 @ s) )
965
+ ( "Rostropovich" n (r o1 s t r @ p ou1 v i ch) )
966
+ ( "Roseanne" n (r ou z a1 n) )
967
+ ( "Peres" n (p e r e1 z) )
968
+ ( "Notre" n (n o1 t r @ ) )
969
+ ( "Mantegna" n (m a n t e g n @) )
970
+ ( "Ligachev" n (l i1 g @ ch @ v) )
971
+ ( "Kurdistan" n (k @@1 d i s t a n) )
972
+ ( "Hernandez" n (h @@ n a1 n d @ z) )
973
+ ( "Genscher" n (g e1 n sh @ r) )
974
+ ( "Diane" n (d ai a1 n) )
975
+ ( "buyouts" n (b ai1 au t s) )
976
+ ( "Yamaguchi" n (y a1 m @ g uu1 ch ii) )
977
+ ( "Wilde" n (w ai1 l d) )
978
+ ( "Stroessner" n (s t r o1 s n @ r) )
979
+ ( "Sakharov" n (s a1 k @ r @ v) )
980
+ ( "MacDougall" n (m @ k d uu1 g @ l) )
981
+ ( "Akihito" n (a1 k ii h ii1 t ou) )
982
+ ( "Yanayev" n (y a1 n @ y @ v) )
983
+ ( "Violeta" n (v ai @ l e1 t @) )
984
+ ( "McDowell" n (m @ k d au1 @ l) )
985
+ ( "Ito" n (ii1 t ou) )
986
+ ( "Dali" n (d a1 l ii) )
987
+ ( "savor" n (s ei1 v @ r) )
988
+ ( "ouster" n (au1 s t @ r) )
989
+ ( "hometown" n (h ou1 m t au n) )
990
+ ( "estrogen" n (e1 s t r @ jh @ n) )
991
+ ( "debuting" n (d ei b y uu1 t i ng) )
992
+ ( "au" n (ou1) )
993
+ ( "Yuri" n (y uu1 r ii) )
994
+ ( "Yakovlev" n (y a1 k @ v l e v) )
995
+ ( "Venturi" n (v e n ch y uu1 r ii) )
996
+ ( "Scorsese" n (s k o r s ei1 s ii) )
997
+ ( "McWhirter" n (m @ k w @@1 t @ r) )
998
+ ( "Kennebunkport" n (k e1 n i b uh1 n k p oo t) )
999
+ ( "Hightower" n (h ai1 t au1 @ r) )
1000
+ ( "Cristiani" n (k r i s t ii a1 n ii) )
1001
+ ( "Crete" n (k r ii1 t) )
1002
+ ( "Ukrainian" n (y uu k r ei1 n i @ n) )
1003
+ ( "Trabi" n (t r a1 b ii) )
1004
+ ( "Schickel" n (sh i1 k @ l) )
1005
+ ( "Miyazawa" n (m i1 y @ z a1 w @) )
1006
+ ( "Gaudi" n (g au1 d ii) )
1007
+ ( "Bastille" n (b a s t ii1 l) )
1008
+ ( "somber" n (s o m b @ r) )
1009
+ ( "prewar" n (p r ii1 w oo1 r) )
1010
+ ( "moviegoers" n (m uu1 v i g ou @ r z) )
1011
+ ( "macho" n (m a1 ch ou) )
1012
+ ( "erstwhile" n (@@ s t w ai l) )
1013
+ ( "cliche" n (k l ii sh ei1) )
1014
+ ( "antiabortion" n (a1 n t i @ b oo1 sh @ n) )
1015
+ ( "Uzbekistan" n (uu z b e1 k i s t a n) )
1016
+ ( "Trabant" n (t r @ b a1 n t ) )
1017
+ ( "Sergei" n (s i1 r g ei) )
1018
+ ( "Montgomery" n (m o n t g o1 m @ r ii) )
1019
+ ( "Ku" n (k uu1) )
1020
+ ( "Heathrow" n (h ii1 th r ou) )
1021
+ ( "Guggenheim" n (g uu1 g @ n h ai m) )
1022
+ ( "Gnehm" n (n e1 m) )
1023
+ ( "Gesell" n (g e1 s @ l) )
1024
+ ( "Dylan" n (d i1 l @ n) )
1025
+ ( "Carolco" n (k a1 r @ l k ou) )
1026
+ ( "Buchwald" n (b u1 k v a l d) )
1027
+ ( "Belorussia" n (b e1 l @ r uh1 sh i @) )
1028
+ ( "kidnaped" n (k i1 d n a1 p t) )
1029
+ ( "Yosemite" n (y @ s e1 m i t ii ) )
1030
+ ( "Ukrainians" n (y uu k r ei1 n i @ n z) )
1031
+ ( "Sylvester" n (s i l v e1 s t @ r) )
1032
+ ( "Schultz" n (sh u1 l t s) )
1033
+ ( "Scharf" n (sh aa1 f) )
1034
+ ( "Reaganomics" n (r ei1 g @ n oo1 m i k s) )
1035
+ ( "Randall" n (r a1 n d @ l) )
1036
+ ( "Puryear" n (p uh1 t y @@ r) )
1037
+ ( "Nidal" n (n i d a1 l) )
1038
+ ( "Nair" n (n e@ r) )
1039
+ ( "Michelangelo" n (m ai1 k @ l a1 n jh @ l ou) )
1040
+ ( "Margot" n (m aa1 g ou) )
1041
+ ( "Leonid" n (l i1 e n @ d) )
1042
+ ( "Lehtinen" n (l ei t i1 n @ n) )
1043
+ ( "Hafez" n (h a1 f e z) )
1044
+ ( "Gomez" n (g ou1 m e z) )
1045
+ ( "Gisela" n (jh i s e1 l @) )
1046
+ ( "Elena" n (e l ei1 n @) )
1047
+ ( "Chi" n (k ai1) )
1048
+ ( "Ariz" n (a1 r i z) )
1049
+ ( "Andre" n (aa1 n d r ei) )
1050
+ ( "prizewinning" n (p r ai1 z w i n i ng) )
1051
+ ( "cliches" n (k l ii sh ei z) )
1052
+ ( "basics" n (b ei1 s i k s) )
1053
+ ( "Yellowstone" n (y e1 l ou s t ou n) )
1054
+ ( "Walesa" n (v a l e1 s @) )
1055
+ ( "Stefan" n (s t e1 f a n) )
1056
+ ( "Schwarzenegger" n (sh w oo1 t z e n ei g @ r) )
1057
+ ( "Sanchez" n (s a1 n ch e z) )
1058
+ ( "Nikita" n (n i k ii1 t @) )
1059
+ ( "Navajo" n (n a1 v @ h ou) )
1060
+ ( "Mitchelson" n (m i1 ch @ l s @ n ) )
1061
+ ( "Menem" n (m e n @ m) )
1062
+ ( "Laurent" n (l oo r o1 n) )
1063
+ ( "Lafester" n (l @ f e1 s t @ r) )
1064
+ ( "Janeiro" n (jh @ n e@1 r ou) )
1065
+ ( "Isaacson" n (ai1 s i k s @ n) )
1066
+ ( "Gromyko" n (g r ou m ii1 k ou) )
1067
+ ( "Fergie" n (f @@1 g ii) )
1068
+ ( "Eco" n (ii1 k ou) )
1069
+ ( "DeWitt" n (d @ w i1 t) )
1070
+ ( "Daryl" n (d a1 r @ l) )
1071
+ ( "policymaking" n (p o1 l i s i m ei k i ng) )
1072
+ ( "policymakers" n (p o1 l i s i m ei k @ r z) )
1073
+ ( "nunciature" n (n @ n s ii1 @ ch u@ r) )
1074
+ ( "noncommunist" n (n o n - k o1 m y uu n i s t) )
1075
+ ( "kidnaping" n (k i d n a1 p i i ng) )
1076
+ ( "facade" n (f @ s aa1 d) )
1077
+ ( "disadvantaged" n (d i s @ d v a1 n t ei jh d) )
1078
+ ( "antigovernment" n (a1 n t i g uh1 v @ n m @ n t) )
1079
+ ( "Yves" n (ii1 v z) )
1080
+ ( "Weisser" n (w ai1 z @ r) )
1081
+ ( "Vitali" n (v i t a1 l ii) )
1082
+ ( "Reiner" n (r ai1 n @ r) )
1083
+ ( "Raisa" n (r ai1 s @) )
1084
+ ( "ROTC" n (aa1 r ou1 t ii1 s ii1) )
1085
+ ( "Nicholson" n (n i1 k @ l s @ n) )
1086
+ ( "Mustafa" n (m uh1 s t a f @) )
1087
+ ( "Gras" n (g r aa1) )
1088
+ ( "Gershwin" n (g @@ sh w @ n) )
1089
+ ( "Gibbs" n (g i1 b z ) )
1090
+ ( "Freud" n (f r oi1 d) )
1091
+ ( "Doubleday" n (d uh1 b @ l d ei) )
1092
+ ( "Beyer" n (b ei1 @ r) )
1093
+ ( "Arrington" n (a1 r i ng t @ n) )
1094
+ ( "Anastasia" n (a1 n a s t ei1 s i @) )
1095
+ ( "Allison" n (a1 l i s o n) )
1096
+ ( "Ahmed" n (aa1 m e d) )
1097
+ ( "geopolitical" n (jh ii1 ou p @ l i1 t @ k @ l) )
1098
+ ( "bookstores" n (b u1 k s t oo1 z) )
1099
+ ( "Zhai" n (z ai1) )
1100
+ ( "Xiaoping" n (sh au1 p i ng) )
1101
+ ( "Xiao" n (sh au1) )
1102
+ ( "Winbush" n (w i1 n b uu sh) )
1103
+ ( "Whitaker" n (w i1 t @ k @ r) )
1104
+ ( "Vytautas" n (v ai t au1 t @ s) )
1105
+ ( "Swaggart" n (s w a1 g @ t) )
1106
+ ( "Reilly" n (r ai1 l ii) )
1107
+ ( "Reid" n (r ii1 d) )
1108
+ ( "Nikolai" n (n i1 k ou l ai) )
1109
+ ( "Motown" n (m ou1 t au n) )
1110
+ ( "Melamed" n (m e1 l @ m e d) )
1111
+ ( "Medved" n (m e1 d v e1 d) )
1112
+ ( "Mazowiecki" n (m @ z au1 s k ii) )
1113
+ ( "Mahon" n (m ei1 h @ n) )
1114
+ ( "Leninist" n (l e1 n @ n i s t) )
1115
+ ( "Jong" n (y o1 ng) )
1116
+ ( "Huston" n (h uu1 s t @ n) )
1117
+ ( "Gumbel" n (g uh1 m b @ l) )
1118
+ ( "Guillermo" n (g u w ii @@1 m ou) )
1119
+ ( "Fuchs" n (f uu1 k s) )
1120
+ ( "Flynn" n (f l i1 n) )
1121
+ ( "Dou" n (d ou1) )
1122
+ ( "Dahmer" n (d aa1 m @ r) )
1123
+ ( "Croatian" n (k r ou ei1 sh @ n) )
1124
+ ( "Capriati" n (k a p r i a1 t ii) )
1125
+ ( "Buthelezi" n (b u t @ l ei1 z ii) )
1126
+ ( "Bhopal" n (b ou p a1 l) )
1127
+ ( "Arcata" n ( aa k a1 t @) )
1128
+ ( "Abrams" n (ei1 b r @ m z) )
1129
+ ( "gutsy" n (g uh1 t s ii) )
1130
+ ( "debuted" n (d ei b y uu1 d) )
1131
+ ( "bipartisanship" n (b ai p aa1 t @ s @ n sh i p) )
1132
+ ( "Webb" n (w e1 b) )
1133
+ ( "Ulam" n (ou1 l @ m) )
1134
+ ( "Shultz" n (sh u1 l t z) )
1135
+ ( "Rosenberg" n (r ou1 s @ n b @@ g) )
1136
+ ( "Regan" n (r ii g @ n) )
1137
+ ( "Noboru" n (n ou b ou1 r uu) )
1138
+ ( "Moshe" n (m o1 sh ei) )
1139
+ ( "Martinez" n (m aa t @ n e1 z) )
1140
+ ( "Klaus" n (k l au1 s) )
1141
+ ( "Islamabad" n (i s l a1 m @ b a d) )
1142
+ ( "Heinrich" n (h ai1 n r i k) )
1143
+ ( "Greenpeace" n (g r ii1 n p ii s) )
1144
+ ( "Geraldo" n (jh e r a1 l d ou) )
1145
+ ( "Donahue" n (d o1 n @ h y uu) )
1146
+ ( "Dmitri" n (d @ m ii1 t r ii) )
1147
+ ( "Dietrich" n (d ii1 t r i k) )
1148
+ ( "Croats" n (k r o1 a t s) )
1149
+ ( "Cione" n (s i ou1 n ii) )
1150
+ ( "Cher" n (sh e1 r) )
1151
+ ( "Chavira" n (sh a v ii1 r @) )
1152
+ ( "Camelot" n (k a1 m @ l o t) )
1153
+ ( "Beethoven" n (b ei1 t h ou v @ n) )
1154
+ ( "Awadi" n (@ w a1 d ii) )
1155
+ ( "Avianca" n (a v i a1 n k @) )
1156
+ ( "Andean" n ( a1 n d i @ n) )
1157
+ ( "Alberto" n (a l b @@1 t ou) )
1158
+ ( "Aaron" n (ei1 r @ n) )
1159
+ ( "misdemeanor" n (m i s d @ m ii1 n @ r ) )
1160
+ ( "mammograms" n (m a1 m @ g r a m z) )
1161
+ ( "detente" n (d ei t o1 n t ) )
1162
+ ( "Vladivostok" n (v l a d @ v o1 s t o k) )
1163
+ ( "Turkmenistan" n (t @@ k m e1 n i s t a n) )
1164
+ ( "Stalinism" n (s t a1 l @ n @ z m) )
1165
+ ( "Seiji" n (s ei1 jh ii) )
1166
+ ( "Prager" n (p r a1 g @ r) )
1167
+ ( "Pinochet" n (p i1 n @ sh ei) )
1168
+ ( "Phnom" n (p @ n o1 m) )
1169
+ ( "Penh" n (p e1 n) )
1170
+ ( "Mengistu" n (m e n g ii1 s t uu) )
1171
+ ( "McManus" n (m @ k m a1 n @ s) )
1172
+ ( "Marsalis" n (m a r s a1 l @ s) )
1173
+ ( "Lou" n (l uu1) )
1174
+ ( "Huang" n (w a1 ng) )
1175
+ ( "Gwynne" n (g w i1 n) )
1176
+ ( "Giroux" n (jh ii r ou1) )
1177
+ ( "Feinstein" n (f ei1 n s t ai n) )
1178
+ ( "Farabundo" n (f a r @ b uh1 n d ou) )
1179
+ ( "Euphrates" n (y uu f r ei1 t ii z) )
1180
+ ( "Dushanbe" n (d @ sh a1 n b @) )
1181
+ ( "Dante" n (d aa1 n t ei) )
1182
+ ( "Chen" n (ch e1 n) )
1183
+ ( "Carr" n (k aa1) )
1184
+ ( "Cadillac" n (k a1 d @ l a k) )
1185
+ ( "Buenos" n (b w ei1 n @ s) )
1186
+ ( "Branagh" n (b r a1 n @) )
1187
+ ( "Bikkembergs" n (b i1 k @ m b @@ g s) )
1188
+ ( "Andreas" n (a n d r ei1 @ s) )
1189
+ ( "wacky" n (w a1 k ii) )
1190
+ ( "stockpiled" n (s t o1 k p ai l d) )
1191
+ ( "salsa" n (s a1 l t s a) )
1192
+ ( "protege" n (p r o1 t @ jh ei) )
1193
+ ( "newsmagazine" n (n y uu1 z m a1 g @ z ii n) )
1194
+ ( "labored" n (l ei1 b @@ d) )
1195
+ ( "deterrence" n (d @ t e@1 r @ n s) )
1196
+ ( "anticommunist" n (a1 n t i k o1 m y uu n @ s t) )
1197
+ ( "acknowledgment" n (@ k n o1 l i jh m @ n t) )
1198
+ ( "Valerie" n (v a1 l @ r ii) )
1199
+ ( "Sheppard" n (sh e1 p @ d) )
1200
+ ( "Ryzhkov" n (r ai1 zh k o f) )
1201
+ ( "Raphael" n (r a f ai1 l) )
1202
+ ( "Quebecois" n (k w i1 b e k w aa) )
1203
+ ( "Profumo" n (p r ou f y u1 m ou) )
1204
+ ( "Mondale" n (m o1 n d ei l) )
1205
+ ( "Minh" n (m i1 n) )
1206
+ ( "Menachem" n ( m e n a1 k @ m) )
1207
+ ( "Mecke" n (m e1 k) )
1208
+ ( "McClure" n (m @ k l u@1) )
1209
+ ( "Magubane" n (m a1 g u b a n ii) )
1210
+ ( "Kupreyev" n (k @ p r ei1 e v) )
1211
+ ( "Krikalev" n (k r i1 k @ l e v) )
1212
+ ( "Khamenei" n (k @ m ei1 n ii) )
1213
+ ( "Kgase" n (k @ g a1 s ii) )
1214
+ ( "Kennan" n (k ii1 n @ n) )
1215
+ ( "Juan" n (w aa1 n ) )
1216
+ ( "Helena" n (h e l e1 n @) )
1217
+ ( "Gomes" n (g ou1 m e z) )
1218
+ ( "Genome" n (jh ii1 n ou m) )
1219
+ ( "Foretich" n (f oo1 r t i k) )
1220
+ ( "Evan" n (e1 v @ n) )
1221
+ ( "Eagleburger" n (ii1 g @ l b @@ g @ r) )
1222
+ ( "Dhabi" n (d aa1 b ii) )
1223
+ ( "Darwin" n (d aa1 w i n) )
1224
+ ( "Camille" n (k a m ii1 l) )
1225
+ ( "Brzezinski" n (b r @ zh i1 n s k ii) )
1226
+ ( "Boyz" n (b oi1 z) )
1227
+ ( "Ayckbourn" n (ei1 k b oo n) )
1228
+ ( "Ariel" n (ei1 r i e l) )
1229
+ ( "Andres" n (a1 n d r e z) )
1230
+ ( "yds" n (y aa1 d z) )
1231
+ ( "sures" n (sh u@1 z) )
1232
+ ( "subtext" n (s uh1 b t e k s t) )
1233
+ ( "siecle" n (s i e1 k @ l ) )
1234
+ ( "savoring" n (s ei1 v @ r i ng) )
1235
+ ( "savior" n (s ei1 v i@ r) )
1236
+ ( "resupply" n (r ii1 s @ p l ai) )
1237
+ ( "paycheck" n (p ei1 ch e k) )
1238
+ ( "ole" n (ou l ei1) )
1239
+ ( "micromastia" n (m ai1 k r o m a1 s t i @) )
1240
+ ( "farmland" n (f aa1 m l a n d) )
1241
+ ( "ethicists" n (e1 th @ s i s t s) )
1242
+ ( "brio" n (b r ii1 ou) )
1243
+ ( "Wynton" n (w i1 n t @ n) )
1244
+ ( "Woodrow" n (w u1 d r ou) )
1245
+ ( "Webber" n (w e1 b @ r) )
1246
+ ( "Uzbeks" n (u1 z b e k s) )
1247
+ ( "Stateside" n (s t ei1 t s ai d) )
1248
+ ( "Smiley" n (s m ai1 l ii) )
1249
+ ( "Schreiber" n (sh r ei1 b @ r) )
1250
+ ( "Ruhollah" n (r u h ou1 l @) )
1251
+ ( "Rivera" n (r i b e1 r @) )
1252
+ ( "Reingold" n (r ai1 n g ou l d) )
1253
+ ( "Purvis" n (p @@1 v i s) )
1254
+ ( "Prawer" n (p r oo1 @ r) )
1255
+ ( "Pamyat" n (p a1 m y a t) )
1256
+ ( "Orion" n (ou r ai1 @ n) )
1257
+ ( "Opel" n (ou1 p @ l) )
1258
+ ( "Oates" n (ou1 t s) )
1259
+ ( "Nina" n (n ii1 n @) )
1260
+ ( "Naisbitt" n (n ei1 z b i t) )
1261
+ ( "NEC" n (e1 n ii1 s ii1) )
1262
+ ( "McElroy" n (m a1 k @ l r oi) )
1263
+ ( "McCrae" n (m @ k r ei1) )
1264
+ ( "Maura" n (m au1 r @) )
1265
+ ( "Maida" n (m ai1 d @) )
1266
+ ( "Lopez" n (l ou1 p e z) )
1267
+ ( "Leonardo" n (l ii o n aa1 d ou) )
1268
+ ( "Kunst" n (k uu1 n s t) )
1269
+ ( "Kouchner" n (k au ch n @ r) )
1270
+ ( "Kiichi" n (k ii ii1 ch ii) )
1271
+ ( "Kagan" n (k ei1 g @ n) )
1272
+ ( "Joaquin" n (jh ou a1 k @ n ) )
1273
+ ( "Jaruzelski" n (y a r @ z e1 l s k ii) )
1274
+ ( "Imelda" n (i m e1 l d @) )
1275
+ ( "Howards" n (h au1 @ d z) )
1276
+ ( "Heisler" n (h ai1 s l @ r) )
1277
+ ( "Guth" n (g u1 th) )
1278
+ ( "Gearheart" n (g i@1 h aa t) )
1279
+ ( "Fukuyama" n (f uu1 k u y a1 m @) )
1280
+ ( "Foucault" n (f ou1 k @ l t) )
1281
+ ( "Farrar" n (f a1 r @ r) )
1282
+ ( "Exocet" n (e1 k s @ s e t) )
1283
+ ( "Etzioni" n (e t z i ou1 n ii) )
1284
+ ( "Entartete" n (e1 n t aa t e t) )
1285
+ ( "Connery" n (k o1 n @ r ii) )
1286
+ ( "Chavez" n (sh a1 v @ z) )
1287
+ ( "Callahan" n (k a1 l @ h a n) )
1288
+ ( "CPOP" n (s ii1 p o p) )
1289
+ ( "Bundestag" n (b u1 n d @ s t a g) )
1290
+ ( "Boyett" n (b oi e1 t) )
1291
+ ( "Bianco" n (b i a1 n k ou) )
1292
+ ( "Axel" n (a1 k s @ l) )
1293
+ ( "Amherst" n (a1 m h @@ s t) )
1294
+ ( "Alfonsin" n (a l f o1 n s @ n) )
1295
+ ( "unsavory" n (uh n s ei1 v r ii) )
1296
+ ( "surefire" n (sh u@1 f ai @) )
1297
+ ( "polygraph" n (p o1 l i g r a f) )
1298
+ ( "nuanced" n (n y uu1 @ n s t) )
1299
+ ( "nisei" n (n ii1 s ei) )
1300
+ ( "mustache" n (m uh s t a1 sh ) )
1301
+ ( "laboring" n (l ei1 b @ r i ng) )
1302
+ ( "laborers" n (l ei1 b @ r @ z) )
1303
+ ( "homegrown" n (h ou1 m g r ou n) )
1304
+ ( "farfetched" n (f aa1 f e ch d) )
1305
+ ( "diva" n (d ii1 v @) )
1306
+ ( "characterizations" n (k a r @ k t @ r ai z ei1 sh @ n) )
1307
+ ( "carcinogen" n (k aa1 s i jh @ n) )
1308
+ ( "behemoth" n (b @ h ii1 m o th) )
1309
+ ( "academe" n (a1 k @ d ii m) )
1310
+
1311
+ ;; More apearing in multiple articles
1312
+
1313
+ ( "Hussein" n (h @ s ei1 n) )
1314
+ ( "perestroika" n (p e r @ s t r oi1 k @) )
1315
+ ( "Tresniowski" n (t r e s n i au1 s k ii) )
1316
+ ( "Monroe" n (m o n r ou1) )
1317
+ ( "Khomeini" n (k o m ei1 n ii) )
1318
+ ( "Lech" n (l e1 k) )
1319
+ ( "Ryan" n (r ai1 @ n) )
1320
+ ( "Fernando" n (f @@ n a1 n d ou) )
1321
+ ( "Gaddafi" n (g @ d a1 f ii) )
1322
+ ( "Muammar" n (m u a1 m @ r) )
1323
+ ( "Melissa" n (m @ l i1 s @) )
1324
+ ( "Cornell" n (k oo n e1 l) )
1325
+ ( "Yevgeni" n (y e v g e1 n ii) )
1326
+ ( "backup" n (b a1 k uh p) )
1327
+ ( "Anatoli" n (a n @ t ou1 l ii) )
1328
+ ( "Tadeusz" n (t a1 d y uu z) )
1329
+ ( "Pablo" n (p a1 b l ou) )
1330
+ ( "Alzheimer" n (a1 l t z ai m @ r) )
1331
+ ( "Alzheimers" n (a1 l t z ai m @ r z) )
1332
+ ( "Calvin" n (k a1 l v @ n) )
1333
+ ( "Bernstein" n (b @@1 n s t ii n) )
1334
+ ( "Ricardo" n (r i k aa1 d ou) )
1335
+ ( "Rafsanjani" n (r a f s @ n jh aa1 n ii) )
1336
+ ( "fatally" n (f ei1 t @ l ii) )
1337
+ ( "cutoff" n (k uh1 t o v) )
1338
+ ( "Alfredo" n (a l f r ei1 d ou) )
1339
+ ( "Alain" n (a l a1 n) )
1340
+ ( "Einstein" n (ai1 n s t ai n) )
1341
+ ( "fi" n (f ii1) )
1342
+ ( "Raul" n (r @ uu1 l) )
1343
+ ( "Bethesda" n (b @ th e1 s d @) )
1344
+ ( "Ungeheuer" n (uh n g @ h oo r) )
1345
+ ( "Rudulph" n (r uu1 d @ l f) )
1346
+ ( "Molotov" n (m o1 l @ t o v) )
1347
+ ( "sci" n (s ai1) )
1348
+ ( "culturally" n (k uh1 l ch @ r @ l ii) )
1349
+ ( "Jaime" n (h ai1 m ii) )
1350
+ ( "Chevy" n (sh e1 v ii) )
1351
+ ( "Chamorro" n (ch @ m o1 r ou) )
1352
+ ( "Auschwitz" n (au1 sh v i t s) )
1353
+ ( "flavored" n (f l ei1 v @ d) )
1354
+ ( "decor" n (d ei k oo1 r) )
1355
+ ( "Olivier" n (@ l i1 v i ei1) )
1356
+ ( "Leona" n (l ii ou1 n @) )
1357
+ ( "spearheaded" n (s p i@1 h e d @ d ) )
1358
+ ( "adamantly" n (a1 d @ m @ n y l ii) )
1359
+ ( "Stallone" n (s t @ l ou1 n) )
1360
+ ( "Ludtke" n (l uh1 d k @) )
1361
+ ( "catalogs" n (k a1 t @ l o g z) )
1362
+ ( "Strauss" n (s t r au1 s) )
1363
+ ( "Schlesinger" n (sh l e1 s i ng @ r) )
1364
+ ( "Rutgers" n (r uh1 t g @ z ) )
1365
+ ( "Georgi" n (g i oo1 g ii) )
1366
+ ( "Faulkner" n (f oo1 l k n @ r) )
1367
+ ( "Bonfante" n (b o1 n f aa n t ii) )
1368
+ ( "plowed" n (p l au1 d) )
1369
+ ( "grownup" n (g r ou1 n uh p) )
1370
+ ( "Zulu" n (z uu1 l uu) )
1371
+ ( "Sri" n (sh r ii1) )
1372
+ ( "Richter" n (r i1 k t @ r) )
1373
+ ( "Oleg" n (ou1 l e g) )
1374
+ ( "Michele" n (m i sh e1 l) )
1375
+ ( "Marcel" n (m aa s e1 l) )
1376
+ ( "Mapplethorpe" n (m ei1 p @ l th oo p) )
1377
+ ( "Cafe" n (k a1 f ei1) )
1378
+ ( "vocals" n (v ou1 k @ l z) )
1379
+ ( "parenting" n (p ei1 r e n t i ng) )
1380
+ ( "emigres" n (e1 m @ g r ei z) )
1381
+ ( "communique" n (k o m y uu1 n i k ei1) )
1382
+ ( "Sunni" n (s uh1 n ii) )
1383
+ ( "Reno" n (r ii1 n ou) )
1384
+ ( "Rajiv" n (r @ jh ii1 v) )
1385
+ ( "Miguel" n (m i g w e1 l) )
1386
+ ( "Malibu" n (m a1 l i b uu) )
1387
+ ( "Inkatha" n (i n k aa1 t @) )
1388
+ ( "HarperCollins" n (h aa1 p @ k oo1 l @ n z) )
1389
+ ( "Ajemian" n (@ jh i1 m i @ n) )
1390
+ ( "superstars" n (s uu1 p @ s t aa z) )
1391
+ ( "rancor" n (r a1 n k oo r) )
1392
+ ( "expansionism" n (e g s p a1 n sh @ n @ z m) )
1393
+ ( "bookstore" n (b u1 k s t oo r) )
1394
+ ( "anemic" n (@ n ii1 m i k) )
1395
+ ( "Rehnquist" n (r e1 n k w i s t) )
1396
+ ( "Rafael" n (r a f ai1 l) )
1397
+ ( "Michaels" n ( m ai1 k @ l z) )
1398
+ ( "Koppel" n (k o1 p @ l) )
1399
+ ( "Crowe" n (k r ou1) )
1400
+ ( "Cesar" n (s ii1 s @ r) )
1401
+ ( "moviemakers" n (m uu1 v i m ei1 k @ r z) )
1402
+ ( "moviegoer" n (m uu1 v i g ou @ r))
1403
+ ( "emigre" n (e1 m i g r ei) )
1404
+ ( "cafes" n (k a f ei1 z) )
1405
+ ( "authored" n (oo1 th @@ d) )
1406
+ ( "Potomac" n (p ou1 t @ m a k) )
1407
+ ( "Javier" n (y ei1 v i ei1) )
1408
+ ( "Eliot" n (e1 l i o t ) )
1409
+ ( "Arsenio" n (aa s ii1 n i ou) )
1410
+ ( "skewed" n (s k y uu1 d) )
1411
+ ( "reinvent" n (r ii1 i n v e1 n t) )
1412
+ ( "naivete" n (n ai ii1 v i t ii) )
1413
+ ( "musings" n (m y uu1 z i ng z) )
1414
+ ( "couture" n (k @ t y u@1) )
1415
+ ( "adrenaline" n (@ d r e1 n @ l i n) )
1416
+ ( "acerbic" n (@ s @@1 b i k) )
1417
+ ( "Mahal" n (m @ h a1 l) )
1418
+ ( "Lauren" n (l o1 r @ n) )
1419
+ ( "Ehrenreich" n (e1 r @ n r i k) )
1420
+ ( "Antonin" n (a1 n t @ n i n) )
1421
+ ( "Anton" n (a1 n t o n) )
1422
+ ( "splashy" n (s p l a1 sh ii) )
1423
+ ( "preschool" n (p r ii1 s k u l) )
1424
+ ( "forebears" n (f oo1 r b e@ z) )
1425
+ ( "Rene" n (r e n ei1) )
1426
+ ( "Katharine" n (k a1 th r i n) )
1427
+ ( "Herzegovina" n (h @@ t s @ g @ v ii1 n @) )
1428
+ ( "Gerhard" n (g e@1 h aa d) )
1429
+ ( "Gerasimov" n (g @ r a1 s @ m o v) )
1430
+ ( "existential" n (e g s i s t e1 n sh @ l) )
1431
+ ( "befuddled" n (b @ f uh1 d @ l d) )
1432
+ ( "Philippe" n (f ii l ii1 p) )
1433
+ ( "PAC" n (p ii1 ei1 s ii1) )
1434
+ ( "Marquez" n (m aa k e1 z) )
1435
+ ( "Kennedys" n (k e1 n @ d @ z) )
1436
+ ( "Jorge" n (y oo jh) )
1437
+ ( "Gallagher" n (g a1 l @ h @ r) )
1438
+ ( "FlorCruz" n (f l oo1 r - k r uu z) )
1439
+ ( "Costello" n (k o s t e1 l ou1) )
1440
+ ( "tenured" n (t e1 n y u@ d) )
1441
+ ( "entrepreneurship" n (o1 n t r @ p r @ n y u@1 sh i p) )
1442
+ ( "elites" n (i l ii1 t s) )
1443
+ ( "Ziyang" n (z ii1 y a ng) )
1444
+ ( "Zhao" n (zh au1) )
1445
+ ( "Leninism" n (l e1 n @ n @ z m) )
1446
+ ( "Lakers" n (l ei1 k @ z) )
1447
+ ( "Kryuchkov" n (k r ai1 ch k o v) )
1448
+ ( "Josef" n (y ou1 s i f) )
1449
+ ( "Gulag" n (g uu1 l a g) )
1450
+ ( "Crimea" n (k r ai m ii1 @) )
1451
+ ( "Caltech" n (k a1 l t e k) )
1452
+ ( "bombings" n (b o1 m i ng z) )
1453
+ ( "Aires" n (ei1 r ii z) )
1454
+ ( "vu" n (v uu1) )
1455
+ ( "viral" n (v ai1 r @ l) )
1456
+ ( "stressful" n (s t r e1 s f @ l) )
1457
+ ( "reprise" n (r i p r ai1 z) )
1458
+ ( "replicate" n (r e1 p l i k ei t) )
1459
+ ( "omnipresent" n (o1 m n i p r e1 s @ n t) )
1460
+ ( "nouveau" n (m uu v ou1 ) )
1461
+ ( "moviemaking" n (m uu1 v i m ei1 k i ng) )
1462
+ ( "hallway" n (h oo1 l w ei1) )
1463
+ ( "fiance" n (f i o1 n s i1) )
1464
+ ( "demographics" n (d e1 m @ g r a1 f i k z) )
1465
+ ( "attache" n (@ t a sh ei1) )
1466
+ ( "Toni" n (t ou1 n ii) )
1467
+ ( "Shimon" n (sh i m o1 n) )
1468
+ ( "Schmidt" n (sh m i1 t) )
1469
+ ( "Revlon" n (r e1 v l o n) )
1470
+ ( "Pont" n (p o1 n t) )
1471
+ ( "Ozal" n (o z a1 l) )
1472
+ ( "Hubble" n (h uh1 b @ l) )
1473
+ ( "Hasan" n (h @ s a1 n) )
1474
+ ( "Fuhrer" n (f y uu1 r @ r) )
1475
+ ( "Aleksandr" n (a l e k z a1 n d @ r) )
1476
+ ( "stanch" n (s t oo1 n ch) )
1477
+ ( "premieres" n (p r e m i @@1 z) )
1478
+ ( "plow" n (p l au1) )
1479
+ ( "passe" n (p a s ei1) )
1480
+ ( "northward" n (n oo1 th w @@ d) )
1481
+ ( "favoritism" n (f ei1 v r @ t @ z m) )
1482
+ ( "deja" n (d ei1 zh @) )
1483
+ ( "caliber" n (k a1 l @ b @ r) )
1484
+ ( "ballyhooed" n (b a l i h uu1 d) )
1485
+ ( "anemia" n (@ n ii1 m i @) )
1486
+ ( "Weimar" n (w ei1 m @ r) )
1487
+ ( "Streisand" n (s t r ai1 s @ n d) )
1488
+ ( "Shakespearean" n (sh ei1 k s p i@ r i @ n) )
1489
+ ( "Seiichi" n (s ei1 ii ch ii) )
1490
+ ( "Sarandon" n (s @ r a1 n d @ n) )
1491
+ ( "Rahman" n (r aa1 m a n) )
1492
+ ( "Manet" n (m a n ei1) )
1493
+ ( "Lamar" n (l @ m aa) )
1494
+ ( "Gup" n (g uh1 p) )
1495
+ ( "Gazeta" n (g @ z e1 t @) )
1496
+ ( "videocassettes" n (v i1 d i ou - k a s e1 t s) )
1497
+ ( "prizewinner" n (p r ai1 z w i n @ r) )
1498
+ ( "nationhood" n (n ei1 sh @ n h uu d) )
1499
+ ( "labors" n (l ei1 b @ z) )
1500
+ ( "hallways" n (h oo1 l w ei z) )
1501
+ ( "genocidal" n (jh e1 n @ s ai1 d @ l) )
1502
+ ( "fide" n (f ai1 d ii) )
1503
+ ( "eponymous" n (e p o1 n @ m @ s) )
1504
+ ( "Zbigniew" n (z @ b i1 g n i e v) )
1505
+ ( "Tito" n (t ii1 t ou) )
1506
+ ( "Schroeder" n (sh r ou1 d @ r) )
1507
+ ( "Romeo" n (r ou1 m i ou) )
1508
+ ( "Roemer" n (r ou1 m @ r) )
1509
+ ( "Mussolini" n (m u1 s i l ii1 n ii) )
1510
+ ( "Lai" n (l ai1) )
1511
+ ( "Karim" n (k a r ii1 m) )
1512
+ ( "Indochina" n (i1 n d @ ch ai1 n @) )
1513
+ ( "Francesco" n (f r a n ch e1 s k ou) )
1514
+ ( "Emilio" n (e m ii1 l i ou) )
1515
+ ( "Ellington" n (e1 l i ng t @ n) )
1516
+ ( "Eisner" n (ai1 s n @ r) )
1517
+ ( "Chopin" n (sh ou1 p a n) )
1518
+ ( "Bowie" n (b au1 ii) )
1519
+ ( "yd" n (y aa1 d) )
1520
+ ( "suburbanites" n (s @ b @@1 b @ n ai1 t z ) )
1521
+ ( "rehab" n (r ii1 h a b) )
1522
+ ( "passersby" n (p a1 s @ s b ai) )
1523
+ ( "officeholders" n (o1 f i s h ou l d @ z) )
1524
+ ( "nomenklatura" n (n o m e1 n k l @ t y uu1 r @) )
1525
+ ( "noir" n (n w aa1 r) )
1526
+ ( "mater" n (m ei1 t @ r) )
1527
+ ( "expansionist" n (i g s p a1 n sh @ n @ s t) )
1528
+ ( "epochal" n (e p o1 k @ l) )
1529
+ ( "entree" n (o1 n t r ei) )
1530
+ ( "denouement" n (d @ n oi1 m e n t) )
1531
+ ( "cutthroat" n (k uh1 t th r ou t) )
1532
+ ( "Versailles" n (v @@ s ai1) )
1533
+ ( "Valeri" n (v a1 l @ r ii) )
1534
+ ( "Su" n (s uu1) )
1535
+ ( "Sharif" n (sh @ r ii1 f) )
1536
+ ( "Saavedra" n (s aa1 v e d r @) )
1537
+ ( "Raleigh" n (r a1 l ei) )
1538
+ ( "Ochoa" n (o ch o1 @) )
1539
+ ( "Marianne" n (m a r i a1 n) )
1540
+ ( "Lucille" n (l uu s ii1 l) )
1541
+ ( "Lindbergh" n (l i1 n d b @@ g) )
1542
+ ( "Hermes" n (h @@1 m ii z) )
1543
+ ( "Goethe" n (g ou1 th @) )
1544
+ ( "Danielle" n (d a n i e1 l) )
1545
+ ( "Constantine" n (k o1 n s t @ n t ii n) )
1546
+ ( "Capone" n (k @ p ou1 n) )
1547
+ ( "Buffett" n (b uu f ei1) )
1548
+ ( "Brigham" n (b r i1 g @ m) )
1549
+ ( "Archie" n (aa1 ch ii) )
1550
+ ( "Amos" n (ei1 m @ s) )
1551
+ ( "Afro" n (a1 f r ou) )
1552
+ ( "unrivaled" n (uh n r ai1 v @ l d) )
1553
+ ( "triage" n (t r ii a1 zh) )
1554
+ ( "theatergoers" n (th i@1 t @ g ou @ r z) )
1555
+ ( "submachine" n (s uh1 b m @ sh ii1 n) )
1556
+ ( "stereotypical" n (s t e1 r i@ t i1 p @ k @ l) )
1557
+ ( "skeletal" n (s k e1 l @ t @ l) )
1558
+ ( "sidekick" n (s ai1 d k i k) )
1559
+ ( "savored" n (s ei1 v @ d) )
1560
+ ( "sadomasochistic" n (s ei1 d ou m a1 s @ k i1 s t i k) )
1561
+ ( "ress" n (r e1 s) )
1562
+ ( "predawn" n (p r ii1 d oo n) )
1563
+ ( "postmodern" n (p ou1 s t m o d @ n) )
1564
+ ( "postcommunist" n (p ou1 s t k o1 m uu n @ s t) )
1565
+ ( "outmaneuvered" n (au1 t m a n uu1 v @@ d) )
1566
+ ( "multicolored" n (m uh1 l t i k uh l @ d) )
1567
+ ( "maneuvered" n (m a n uu1 v @@ d) )
1568
+ ( "liter" n (l ii1 t @ r) )
1569
+ ( "laborer" n (l ei1 b @ r @ r) )
1570
+ ( "falloff" n (f oo1 l o v) )
1571
+ ( "facades" n (f @ s a1 d z) )
1572
+ ( "destabilize" n (d ii s t ei1 b @ l ai z) )
1573
+ ( "coincidentally" n (k ou1 i n s @ d e1 n t @ l ii) )
1574
+ ( "carcinogens" n (k aa1 s i jh @ n z) )
1575
+ ( "antinuclear" n (a1 n t i n y uu1 k l i@ r) )
1576
+ ( "aerobic" n (ei r ou1 b i k ) )
1577
+ ( "Yusuf" n (y uu1 s @ f) )
1578
+ ( "Willem" n (w i1 l e m) )
1579
+ ( "Ventura" n (v e n ch y uu1 r @) )
1580
+ ( "Saharan" n (s @ h a1 r @ n) )
1581
+ ( "Rutherford" n (r uh1 th @ f oo d) )
1582
+ ( "Psycho" n (s ai1 k ou) )
1583
+ ( "Miserables" n (m ii z @ r aa1 b l @) )
1584
+ ( "Mimi" n (m ii1 m ii) )
1585
+ ( "Milos" n (m ai1 l ou s) )
1586
+ ( "Magellan" n (m a g e1 l @ n) )
1587
+ ( "Kosovo" n (k @ s ou1 v ou) )
1588
+ ( "Komsomolskaya" n (k o1 m s @ m o1 l s k ai @) )
1589
+ ( "Julio" n (jh uu1 l i ou) )
1590
+ ( "Jozef" n (y ou1 z e f) )
1591
+ ( "Gucci" n (g uu1 ch ii) )
1592
+ ( "Gonzalo" n (g o n z a1 l ou) )
1593
+ ( "Gillespie" n (g i l e1 s p ii) )
1594
+ ( "Ghaith" n (g ei1 th) )
1595
+ ( "Gacha" n (g a1 ch @) )
1596
+ ( "Fairbanks" n (f e@1 t b a n k s) )
1597
+ ( "Duvalier" n (d u v a1 l i ei) )
1598
+ ( "Dont" n (d ou1 n t) )
1599
+ ( "Danube" n (d a1 n y uu b) )
1600
+ ( "Cong" n (k o1 ng) )
1601
+ ( "Cajun" n (k ei1 jh @ n) )
1602
+ ( "Angelenos" n (a1 n jh i l ii1 n ou s) )
1603
+ ( "Abel" n (ei1 b @ l) )
1604
+ ( "Allis" n (a1 l @ s) )
1605
+ ( "zeitgeist" n (z ai1 t g ai s t) )
1606
+ ( "superhero" n (s uu1 p @ h ii1 r ou) )
1607
+ ( "racetrack" n (r ei1 s - t r a k) )
1608
+ ( "materiel" n (m a t ii1 r i e l) )
1609
+ ( "jazzman" n (jh a1 z m a n) )
1610
+ ( "greats" n (g r ei1 t s) )
1611
+ ( "faux" n (f ou1) )
1612
+ ( "emphysema" n (e1 m f @ s ii1 m @) )
1613
+ ( "divisiveness" n (d @ v ai1 s i f n @ s) )
1614
+ ( "bicoastal" n (b ai1 k ou1 s t @ l) )
1615
+ ( "apolitical" n (ei1 p @ l i1 t @ k @ l) )
1616
+ ( "Shepard" n (sh e1 p @ d) )
1617
+ ( "Romero" n (r ou m e1 r ou) )
1618
+ ( "Pacino" n (p @ ch ii1 n ou) )
1619
+ ( "Montenegro" n (m o n t @ n ii1 g r ou) )
1620
+ ( "Lourdes" n (l uu1 r @ d z) )
1621
+ ( "Kai" n (k ai1) )
1622
+ ( "Jamil" n (jh a m ii1 l ) )
1623
+ ( "Horatio" n (h @ r ei1 sh i ou) )
1624
+ ( "Elias" n (e1 l i @ s) )
1625
+ ( "Dumas" n (d uu1 m a s) )
1626
+ ( "Delors" n (d @ l oo1 r z) )
1627
+ ( "Chang" n (ch a1 ng) )
1628
+ ( "Andropov" n (a n d r o1 p o v ) )
1629
+ ( "Alvarez" n (a l v a1 r e z) )
1630
+ ( "whammy" n (w a1 m ii) )
1631
+ ( "wannabes" n (w o1 n @ b ii z) )
1632
+ ( "underage" n (uh1 n d @ r ei1 jh) )
1633
+ ( "uncaring" n (uh n k ei1 r i ng) )
1634
+ ( "sidelined" n (s ai1 d l ai n d) )
1635
+ ( "saber" n (s ei1 b @ r) )
1636
+ ( "redraw" n (r ii1 d r oo) )
1637
+ ( "recyclable" n (r ii s ai1 k l @ b @ l) )
1638
+ ( "phenom" n (f ii1 n ou m) )
1639
+ ( "paternalistic" n (p a t @@1 n @ l i s t i k) )
1640
+ ( "overleveraged" n (ou1 v @ l e1 v @ r ei jh d) )
1641
+ ( "nouvelle" n (n uu v e1 l) )
1642
+ ( "microchip" n (m ai1 k r ou ch i p) )
1643
+ ( "lifesaving" n (l ai1 f s ei1 v i ng) )
1644
+ ( "hormonal" n (h oo r m ou1 n @ l) )
1645
+ ( "homophobic" n (h ou1 m @ f ou1 b i k) )
1646
+ ( "gravitational" n (g r a v @ t ei1 sh @ n @ l) )
1647
+ ( "faxes" n (f a1 k s @ s) )
1648
+ ( "energize" n (e1 n @ jh ai z) )
1649
+ ( "counterrevolution" n (k au1 n t @ r e1 v @ l uu1 sh @ n) )
1650
+ ( "burgers" n (b @@1 g @ z) )
1651
+ ( "bimonthly" n (b ai1 m uh1 n th l ii) )
1652
+ ( "Zulus" n (z uu1 l uu z) )
1653
+ ( "Zimmerman" n (z i1 m @ m @ n) )
1654
+ ( "Zia" n (z ii1 @) )
1655
+ ( "Vincennes" n (v a1 n s e n) )
1656
+ ( "Tri" n (t r ai1) )
1657
+ ( "Tigris" n (t ai1 g r i s) )
1658
+ ( "Schumer" n (sh u1 m @ r) )
1659
+ ( "Prokofiev" n (p r ou k o1 f i e v) )
1660
+ ( "Minnelli" n (m i n e1 l ii) )
1661
+ ( "McMurtry" n (m @ k m @@1 t ii) )
1662
+ ( "Macbeth" n (m a k b e1 th) )
1663
+ ( "Loren" n (l oo1 r @ n) )
1664
+ ( "Levi" n (l ii1 v ai) )
1665
+ ( "Kumiko" n (k uu1 m i k ou) )
1666
+ ( "Jovanovich" n (jh ou v a1 n @ v i ch) )
1667
+ ( "Jamieson" n (jh ei1 m @ s @ n) )
1668
+ ( "Guthrie" n (g uh1 th r ii) )
1669
+ ( "Frankenstein" n (f r a1 n k @ n s t ai n) )
1670
+ ( "Farrakhan" n (f a1 r @ k aa n) )
1671
+ ( "Dubai" n (d uu b ai1) )
1672
+ ( "Dreyfuss" n (d r ei1 f @ s) )
1673
+ ( "Croatians" n (k r ou ei1 sh @ n z) )
1674
+ ( "Cezanne" n (s e z a1 n) )
1675
+ ( "Basinger" n (b a1 s i n jh @ r) )
1676
+ ( "Bankshares" n (b a1 n k sh e@ z) )
1677
+ ( "Antoine" n (a1 n t w o n) )
1678
+ ( "Achille" n (a k i1 l ii) )
1679
+ ( "trucking" n (t r uh1 k i ng) )
1680
+ ( "storefronts" n (s t oo1 r f r uh n t s) )
1681
+ ( "socioeconomic" n (s ou1 sh i o e1 k @ n o1 m i k) )
1682
+ ( "rivaled" n (r ai1 v @ l d) )
1683
+ ( "reusable" n (r ii y uu1 z @ b @ l) )
1684
+ ( "pizazz" n (p i z a1 z) )
1685
+ ( "overreaction" n (ou1 v @ r ii a1 k sh @ n) )
1686
+ ( "osteoporosis" n (o1 s t i o p o r ou1 s @ s) )
1687
+ ( "limo" n (l i1 m ou) )
1688
+ ( "interagency" n (i1 n t @ ei1 jh e n s ii) )
1689
+ ( "imperatives" n (i m p e1 r @ t i v z) )
1690
+ ( "humorless" n (h y uu1 m @ l @ s) )
1691
+ ( "flatbed" n (f l a1 t b e d) )
1692
+ ( "extraterrestrial" n (e1 k s t r @ t @ r e1 s t r i @ l) )
1693
+ ( "colorless" n (k uh1 l @ r l @ s) )
1694
+ ( "coed" n (k ou e1 d) )
1695
+ ( "chauffeured" n (sh ou f @@1 d) )
1696
+ ( "bluesy" n (b l uu1 z ii) )
1697
+ ( "biomedical" n (b ai1 ou m e1 d i k @ l) )
1698
+ ( "biochemist" n (b ai1 ou k e1 m @ s t ) )
1699
+ ( "bebop" n (b ii1 b o p) )
1700
+ ( "barroom" n (b aa1 r uu m) )
1701
+ ( "accompli" n (@ k o1 m p l ei) )
1702
+ ( "Vaughan" n (v oo1 n) )
1703
+ ( "Ty" n (t ai1) )
1704
+ ( "Szczesny" n (zh @ zh e1 s n ii) )
CosyVoice-ttsfrd/resource/festival/dicts/oald/oald_lts_rules.scm ADDED
The diff for this file is too large to render. See raw diff
 
CosyVoice-ttsfrd/resource/festival/dicts/oald/oaldlex.scm ADDED
@@ -0,0 +1,347 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1998 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+
34
+
35
+ (defvar oaldlexdir (path-append lexdir "oald"))
36
+
37
+ (require 'pos)
38
+
39
+ (define (oald_lts_function word feats)
40
+ "(oald_lts_function word feats)
41
+ Function called when word is not found in lexicon. Uses
42
+ LTS rules trained from the original lexicon, and lexical stress
43
+ prediction rules."
44
+ (require 'lts)
45
+ (if (not (boundp 'oald_lts_rules))
46
+ (load (path-append oaldlexdir "oald_lts_rules.scm")))
47
+ (let ((dcword (downcase word))
48
+ (syls) (phones))
49
+ (if (string-matches dcword "[a-z]*")
50
+ (begin
51
+ (set! lts_pos feats)
52
+ (set! phones (lts_predict dcword oald_lts_rules))
53
+ (set! syls (lex.syllabify.phstress phones))
54
+ )
55
+ (set! syls nil))
56
+ (list word nil syls)))
57
+
58
+ (define (mrpa_addenda)
59
+ "(mrpa_addenda)
60
+ Add a whole host of various entries to the current lexicon with
61
+ mrpa phones."
62
+ ; (lex.add.entry
63
+ ; '("a's" dt (((ei z) 1))))
64
+ ; (lex.add.entry
65
+ ; '("e's" dt (((ii z) 1))))
66
+ ; (lex.add.entry
67
+ ; '("i's" dt (((ai z) 1))))
68
+ ; (lex.add.entry
69
+ ; '("o's" dt (((ou z) 1))))
70
+ ; (lex.add.entry
71
+ ; '("u's" dt (((y uu z) 1))))
72
+ ; (lex.add.entry
73
+ ; '("r's" dt (((aa z) 1))))
74
+ (lex.add.entry
75
+ '("a" dt (((@) 0))))
76
+ (lex.add.entry
77
+ '("a" n (((ei) 1))))
78
+ (lex.add.entry
79
+ '("ac" n (((ei) 1) ((s ii) 1))))
80
+ (lex.add.entry
81
+ '("us" prp (((uh s) 0))))
82
+ (lex.add.entry
83
+ '("'s" pos (((@ z) 0))))
84
+ (lex.add.entry
85
+ '("'s" n (((@ z) 0))))
86
+ (lex.add.entry
87
+ '("the" dt (((dh @) 0))))
88
+ (lex.add.entry
89
+ '("taylor" n (((t ei) 1) ((l @) 0))))
90
+ (lex.add.entry
91
+ '("who" prp ((( h uu ) 0))))
92
+ (lex.add.entry
93
+ '("algorithm" n (((a l) 1) ((g o) 0) ((r i th m) 0))))
94
+ (lex.add.entry
95
+ '("algorithms" n (((a l) 1) ((g o) 0) ((r i th m z) 0))))
96
+ (lex.add.entry
97
+ '("algorithmic" n (((a l) 2) ((g o) 0) ((r i th) 1) ((m i k) 0))))
98
+ (lex.add.entry
99
+ '("alices" n (((a) 1) ((l i) 0) ((s i z) 0))))
100
+ (lex.add.entry
101
+ '( "Angeles" n (((a n) 1) ((jh i) 0) ((l ii z) 0))))
102
+ (lex.add.entry
103
+ '( "atr" n ((( ei ) 1) (( t ii ) 1) (( aa ) 1))))
104
+ (lex.add.entry
105
+ '( "att" n ((( ei ) 1) (( t ii ) 1) (( a n d ) 0) (( t ii ) 1))))
106
+ (lex.add.entry
107
+ '( "awb" n ((( ei ) 1) ((d uh b) 1) ((l y uu) 0) ((b ii) 1))))
108
+ (lex.add.entry
109
+ '("color" n (((k uh) 1) ((l @) 0))))
110
+ (lex.add.entry
111
+ '("colors" n (((k uh) 1) ((l @ z) 0))))
112
+ (lex.add.entry
113
+ '("colored" j (((k uh) 1) ((l @ d) 0))))
114
+ (lex.add.entry
115
+ '("coloring" j (((k uh) 1) ((l @) 0) ((r i ng ) 0))))
116
+ (lex.add.entry
117
+ '("cdrom" n (((s ii) 1) ((d ii) 1) ((r o m) 1))))
118
+ (lex.add.entry
119
+ '("cdroms" n (((s ii) 1) ((d ii) 1) ((r o m z) 1))))
120
+ (lex.add.entry
121
+ '("cepstra" n (((k e p) 1) ((s t r @) 0))))
122
+ (lex.add.entry
123
+ '("cepstral" n (((k e p) 1) ((s t r @ l) 0))))
124
+ (lex.add.entry
125
+ '("cepstrum" n (((k e p) 1) ((s t r @ m) 0))))
126
+ (lex.add.entry
127
+ '("co" nil (((k ou) 1))))
128
+ (lex.add.entry
129
+ '( "cstr" n ((( s ii ) 1) (( e s ) 1) (( t ii ) 1) (( aa ) 1)) ))
130
+ (lex.add.entry
131
+ '( "cogsci" n ((( k o g ) 1) (( s ai) 1))))
132
+ (lex.add.entry
133
+ '("database" n (((d ei) 1) ((t @) 0) ((b ei s) 1))))
134
+ (lex.add.entry
135
+ '("databases" n (((d ei) 1) ((t @) 0) ((b ei) 1) ((s i z) 0))))
136
+ (lex.add.entry
137
+ '("diphone" n (((d ai) 1) ((f ou n) 0))))
138
+ (lex.add.entry
139
+ '("diphones" n (((d ai) 1) ((f ou n s) 0))))
140
+ (lex.add.entry
141
+ '( "edinburgh" n ((( e ) 1) ((d i n ) 0) ((b r @) 0))))
142
+ (lex.add.entry
143
+ '( "email" n ((( ii ) 1) (( m ei l) 0))))
144
+ (lex.add.entry
145
+ '( "emailed" n ((( ii ) 1) (( m ei l d) 0))))
146
+ (lex.add.entry
147
+ '( "emacs" n ((( ii ) 1) (( m a k s) 0))))
148
+ (lex.add.entry
149
+ '( "favorite" j (((f ei) 1) ((v @) 0) ((r i t) 0))))
150
+ (lex.add.entry
151
+ '( "favor" n (((f ei) 1) ((v @) 0))))
152
+ (lex.add.entry
153
+ '( "favors" n (((f ei) 1) ((v @ z) 0))))
154
+ (lex.add.entry
155
+ '( "favoring" n (((f ei) 1) ((v @) 0) ((r i ng) 0))))
156
+ (lex.add.entry
157
+ '( "favored" n (((f ei) 1) ((v @ d) 0))))
158
+ (lex.add.entry
159
+ '("globally" a (((g l ou) 1) ((b @) 0) ((l ii) 0))))
160
+ (lex.add.entry
161
+ '("gorbachev" m (((g oo) 1) ((b @) 0) ((ch e v) 0))))
162
+ (lex.add.entry
163
+ '("grave" n (((g r ei v) 1))))
164
+ (lex.add.entry
165
+ '("graves" j (((g r ei v z) 1))))
166
+ (lex.add.entry
167
+ '("greece" n (((g r ii s) 1))))
168
+ (lex.add.entry
169
+ '("hong" j (((h o ng) 1))))
170
+ (lex.add.entry
171
+ '("hz" n (((h @@ t z) 1))))
172
+ (lex.add.entry
173
+ '("lead" led (((l e d) 1))))
174
+ (lex.add.entry
175
+ '("lead" liid (((l ii d) 1))))
176
+ (lex.add.entry
177
+ '("innovative" j (((i) 2) ((n @) 0) ((v ei) 1) ((t i v) 0))))
178
+ (lex.add.entry
179
+ '("job" n (((jh o b) 1))))
180
+ (lex.add.entry
181
+ '("jobs" n (((jh o b z) 1))))
182
+ (lex.add.entry
183
+ '( "Jr" n (((jh uu) 1) ((n i@) 0))))
184
+ (lex.add.entry
185
+ '("kong" n (((k o ng) 1))))
186
+ (lex.add.entry
187
+ '("khz" n (((k i) 1) ((l ou) 0) ((h @@ t z) 1))))
188
+ (lex.add.entry
189
+ '("labor" n (((l ei) 1) ((b @) 0))))
190
+ (lex.add.entry
191
+ '( "Los" n ((( l o s) 1))))
192
+ (lex.add.entry
193
+ '("lower" v (((l ou) 1) ((@) 0))))
194
+ (lex.add.entry
195
+ '("lowered" v (((l ou) 1) ((@ d) 0))))
196
+ (lex.add.entry
197
+ '("lowering" v (((l ou) 1) ((@) 0) ((r i ng ) 0))))
198
+ (lex.add.entry
199
+ '( "mbrola" n (((e m) 0) ((b r ou) 1) ((l @ ) 0))))
200
+ (lex.add.entry
201
+ '("mhz" n (((m e) 1) ((g @) 0) ((h @@ t z) 1))))
202
+ (lex.add.entry
203
+ '("minute" n (((m i) 1) ((n i t) 0))))
204
+ (lex.add.entry
205
+ '("ms" n (((e m) 1) ((e s) 1))))
206
+ (lex.add.entry
207
+ '("no" dt (((n ou) 1))))
208
+ (lex.add.entry
209
+ '("pault" n ((( p oo l) 1) ((t ii) 1))))
210
+ (lex.add.entry
211
+ '("put" v (((p u t) 1))))
212
+ (lex.add.entry
213
+ '("putting" v (((p u) 1) ((t i ng) 0))))
214
+ (lex.add.entry
215
+ '( "psola" n ((( p i ) 0) (( s ou ) 1) (( l @ ) 0 ))))
216
+ (lex.add.entry
217
+ '("read" red (((r e d) 1))))
218
+ (lex.add.entry
219
+ '("read" riid (((r ii d) 1))))
220
+ (lex.add.entry
221
+ '("reuter" n (((r oi) 1) ((t @@) 0))))
222
+ (lex.add.entry
223
+ '("reuters" n (((r oi) 1) ((t @@ s) 0))))
224
+ (lex.add.entry
225
+ '("row" v (((r ou) 1))))
226
+ (lex.add.entry
227
+ '("row" n (((r ou) 1))))
228
+ (lex.add.entry
229
+ '( "San" n (((s a n) 1))))
230
+ (lex.add.entry
231
+ '("second" n (((s e) 1) ((k @ n d) 0))))
232
+ (lex.add.entry
233
+ '("seconds" n (((s e) 1) ((k @ n d z) 0))))
234
+ (lex.add.entry
235
+ '("sixteenth" n (((s i k s) 2) ((t ii n th) 1))))
236
+ (lex.add.entry
237
+ '("sony" n (((s ou) 1) ((n ii) 1))))
238
+ (lex.add.entry
239
+ '("SSML" v (((e s) 1) ((e s) 1) ((e m) 1) ((e l) 0))))
240
+ (lex.add.entry
241
+ '("sun" n (((s uh n) 1))))
242
+ (lex.add.entry
243
+ '("synthesise" v (((s i n) 1) ((th @) 0) ((s ai z) 0))))
244
+ (lex.add.entry
245
+ '("synthesised" v (((s i n) 1) ((th @) 0) ((s ai z d) 0))))
246
+ (lex.add.entry
247
+ '("synthesizer" n (((s i n) 1) ((th @) 0) ((s ai) 0) ((z @) 0))))
248
+ (lex.add.entry
249
+ '("synthesiser" n (((s i n) 1) ((th @) 0) ((s ai) 0) ((z @) 0))))
250
+ (lex.add.entry
251
+ '("synthesizers" n (((s i n) 1) ((th @) 0) ((s ai) 0) ((z @ z) 0))))
252
+ (lex.add.entry
253
+ '("synthesisers" n (((s i n) 1) ((th @) 0) ((s ai) 0) ((z @ z) 0))))
254
+ (lex.add.entry
255
+ '( "and" aend ((( a n d ) 1))))
256
+ (lex.add.entry
257
+ '( "and" ahnd ((( @ n d ) 0))))
258
+ (lex.add.entry
259
+ '( "to" to ((( t @ ) 0))))
260
+ (lex.add.entry
261
+ '( "to" tuu ((( t uu ) 1))))
262
+ (lex.add.entry
263
+ '("for" faor (((f oo) 1))))
264
+ (lex.add.entry
265
+ '("for" fer (((f @) 0))))
266
+ (lex.add.entry
267
+ '( "tobi" n ((( t ou ) 1) (( b ii ) 0))))
268
+ (lex.add.entry
269
+ '("todays" n (((t @) 0) ((d ei z) 1))))
270
+ (lex.add.entry
271
+ '( "tts" n ((( t ii ) 1) (( t ii ) 1) (( e s ) 1))))
272
+ (lex.add.entry
273
+ '( "ulaw" n ((( m y uu ) 1) (( l oo ) 0))))
274
+ (lex.add.entry
275
+ '( "uk" n ((( y uu ) 1) (( k ei ) 1))))
276
+ (lex.add.entry
277
+ '("waveform" n (((w ei v) 1) ((f oo m) 0))))
278
+ (lex.add.entry
279
+ '("winds" n (((w i n d z) 1)) ))
280
+ (lex.add.entry
281
+ '("wind" v (((w ai n d) 1)) ))
282
+ (lex.add.entry
283
+ '("within" a (((w i) 0) ((dh i n) 1))))
284
+ (lex.add.entry
285
+ '("*" n (((a s) 1) ((t @) 0) ((r i s k) 0))))
286
+ (lex.add.entry
287
+ '("%" n (((p @) 0) ((s e n t) 1))))
288
+ (lex.add.entry
289
+ '("&" n (((a n d) 1)))) ; Original: (((a m p) 1) ((@@ s) 0) ((a n d) 0)))
290
+ (lex.add.entry
291
+ '("$" n (((d o) 1) ((l @) 0))))
292
+ (lex.add.entry
293
+ '("#" n (((h a sh) 1))))
294
+ (lex.add.entry
295
+ '("@" n (((a t) 1))))
296
+ (lex.add.entry
297
+ '("+" n (((p l uh s) 0))))
298
+ (lex.add.entry
299
+ '("^" n (((k a) 1) ((r e t) 0))))
300
+ (lex.add.entry
301
+ '("~" n (((t i l) 1) ((d @) 0))))
302
+ (lex.add.entry
303
+ '("=" n (((ii) 1) ((k w @ l) 0))))
304
+ (lex.add.entry
305
+ '("/" n (((s l a sh) 1))))
306
+ (lex.add.entry
307
+ '("\\" n (((b a k) 1) ((s l a sh) 1))))
308
+ (lex.add.entry
309
+ '("_" n (((uh n) 1) ((d @) 0) ((s k oo) 1))))
310
+ (lex.add.entry
311
+ '("|" n (((v @@) 1) ((t i) 0) ((k @ l) 0) ((b aa) 1))))
312
+ (lex.add.entry
313
+ '(">" n ((( g r ei ) 1) ((t @@) 0) ((dh a n) 1))))
314
+ (lex.add.entry
315
+ '("<" n ((( l e s) 1) ((dh a n) 1))))
316
+ (lex.add.entry
317
+ '("[" n ((( l e f t) 1) ((b r a) 1) ((k i t) 0))))
318
+ (lex.add.entry
319
+ '("]" n ((( r ai t) 1) ((b r a) 1) ((k i t) 0))))
320
+ (lex.add.entry
321
+ '(" " n (((s p ei s) 1))))
322
+ (lex.add.entry
323
+ '("\t" n (((t a b) 1))))
324
+ (lex.add.entry
325
+ '("\n" n (((n y uu) 1) ((l ai n) 1))))
326
+ (lex.add.entry '("." punc nil))
327
+ (lex.add.entry '("'" punc nil))
328
+ (lex.add.entry '(":" punc nil))
329
+ (lex.add.entry '(";" punc nil))
330
+ (lex.add.entry '("," punc nil))
331
+ (lex.add.entry '("-" punc nil))
332
+ (lex.add.entry '("\"" punc nil))
333
+ (lex.add.entry '("`" punc nil))
334
+ (lex.add.entry '("?" punc nil))
335
+ (lex.add.entry '("!" punc nil))
336
+ )
337
+
338
+ (lex.create "oald")
339
+ (lex.set.compile.file (path-append oaldlexdir "ali.oald.lex"))
340
+ (lex.set.phoneset "mrpa")
341
+ (lex.set.lts.method 'oald_lts_function)
342
+ (lex.set.pos.map english_pos_map_wp39_to_wp20)
343
+ (mrpa_addenda)
344
+
345
+
346
+ (provide 'oaldlex)
347
+
CosyVoice-ttsfrd/resource/festival/dicts/wsj.wp39.poslexR ADDED
The diff for this file is too large to render. See raw diff
 
CosyVoice-ttsfrd/resource/festival/dicts/wsj.wp39.tri.ngrambin ADDED
Binary file (144 kB). View file
 
CosyVoice-ttsfrd/resource/festival/etc/email_filter ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/bin/sh
2
+ ###########################################################################
3
+ ## ##
4
+ ## Centre for Speech Technology Research ##
5
+ ## University of Edinburgh, UK ##
6
+ ## Copyright (c) 1996,1997 ##
7
+ ## All Rights Reserved. ##
8
+ ## ##
9
+ ## Permission is hereby granted, free of charge, to use and distribute ##
10
+ ## this software and its documentation without restriction, including ##
11
+ ## without limitation the rights to use, copy, modify, merge, publish, ##
12
+ ## distribute, sublicense, and/or sell copies of this work, and to ##
13
+ ## permit persons to whom this work is furnished to do so, subject to ##
14
+ ## the following conditions: ##
15
+ ## 1. The code must retain the above copyright notice, this list of ##
16
+ ## conditions and the following disclaimer. ##
17
+ ## 2. Any modifications must be clearly marked as such. ##
18
+ ## 3. Original authors' names are not deleted. ##
19
+ ## 4. The authors' names are not used to endorse or promote products ##
20
+ ## derived from this software without specific prior written ##
21
+ ## permission. ##
22
+ ## ##
23
+ ## THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ##
24
+ ## DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ##
25
+ ## ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ##
26
+ ## SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ##
27
+ ## FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ##
28
+ ## WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ##
29
+ ## AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ##
30
+ ## ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ##
31
+ ## THIS SOFTWARE. ##
32
+ ## ##
33
+ ###########################################################################
34
+ ## ##
35
+ ## Email filter for tts text mode ##
36
+ ## usage: email_filter email_message >filtered_message ##
37
+ ## ##
38
+ ## Extracts the From and Subject lines form the head and the body of ##
39
+ ## of the message, I suppose it could also do signature extraction ##
40
+ ## ##
41
+ ###########################################################################
42
+ grep "^From: " $1
43
+ echo
44
+ grep "^Subject: " $1
45
+ echo
46
+ # delete up to first blank line (i.e. the header)
47
+ sed '1,/^$/ d' $1
CosyVoice-ttsfrd/resource/festival/multisyn/multisyn.scm ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 2003, 2004 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Multisyn top level scheme code (Korin Richmond and Rob Clark)
35
+ ;;;
36
+
37
+ ; Requires
38
+ (require_module 'UniSyn)
39
+ (require_module 'MultiSyn)
40
+ (require 'multisyn_pauses)
41
+ (require 'target_cost)
42
+
43
+ ;; use a global parameter to specify which UnitSelection voice
44
+ ;; to use to synthesise a given utterance for now, because the
45
+ ;; standard Festival synthesis mainline doesn't accept a voice
46
+ ;; parameter. (This should be set to the current voice object)
47
+ (defvar currentMultiSynVoice nil)
48
+ (defvar relp t)
49
+ (defvar flattenVoice nil)
50
+
51
+ ; extract utt list from a .data file
52
+ (define (load_utt_list filename)
53
+ "(load_utt_list filename)
54
+ Loads a fextvox .data file and extracts an utterance list."
55
+ (let (l entries)
56
+ (set! entries (load filename t))
57
+ (mapcar
58
+ (lambda (d)
59
+ (set! l (cons (car d) l))
60
+ t)
61
+ entries)
62
+ l))
63
+
64
+ ;; SynthType definition, main entry point.
65
+
66
+ (defSynthType MultiSyn
67
+ ;(print "Multisyn unit selection synthesis")
68
+ (defvar MultiSyn_module_hooks nil)
69
+ (Param.def "unisyn.window_name" "hanning")
70
+ (Param.def "unisyn.window_factor" 1.0)
71
+ ;; Unisyn requires these to be set.
72
+ (set! us_abs_offset 0.0)
73
+ (set! us_rel_offset 0.0)
74
+
75
+ (apply_hooks MultiSyn_module_hooks utt) ;; 4processing of diphone names
76
+
77
+ ;; find appropriate unit sequence and put sythesis
78
+ ;; parameters in the Unit relation of the utterance structure
79
+ (voice.getUnits currentMultiSynVoice utt)
80
+
81
+ ;(print "doing concat")
82
+ (us_unit_concat utt)
83
+
84
+ ;(print "doing raw concat")
85
+
86
+ (utt.relation.create utt 'SourceSegments)
87
+
88
+ (set! do_prosmod (du_voice.prosodic_modification currentMultiSynVoice))
89
+
90
+ (if do_prosmod
91
+ (begin
92
+ (if (not (member 'f0 (utt.relationnames utt)))
93
+ (targets_to_f0 utt))
94
+ ;; temporary fix
95
+ (if (utt.relation.last utt 'Segment)
96
+ (set! pm_end (+ (item.feat (utt.relation.last utt 'Segment) "end") 0.02))
97
+ (set! pm_end 0.02))
98
+ (us_f0_to_pitchmarks utt 'f0 'TargetCoef pm_end)
99
+ (us_mapping utt 'segment_single))
100
+ (begin
101
+ (utt.copy_relation utt 'SourceCoef 'TargetCoef)
102
+ (us_mapping utt "linear")))
103
+
104
+
105
+ ;(print "generating wave")
106
+ ;; specify something else if you don't want lpc
107
+ (us_generate_wave utt 'lpc)
108
+ )
109
+
110
+
111
+ ; target cost scheme code
112
+ (define (targetcost it1 it2)
113
+ (Default_Target_Cost it1 it2))
114
+
115
+ ; Evil function which writes the functions to actually load and switch new voices.
116
+ (define (make_voice_definition name srate config_function backoff_rules data_dir config)
117
+ "(make_voice_definition NAME SRATE CONFIG_FUNCTION BACKOFF_RULES DATA_DIR CONFIG)
118
+ Create the fuction definitions to load and unload a voice."
119
+ (let ((voice_name (string-append "voice_" name))
120
+ (free_name (string-append "free_voice_" name))
121
+ (pre_config_function (string-append config_function "_pre"))
122
+ (voice_variable (upcase (string-append "voice_" name))))
123
+
124
+ (eval (list 'defvar (intern voice_variable) nil))
125
+
126
+ (eval (list 'define (list (intern voice_name))
127
+ (list 'if (intern pre_config_function)
128
+ (list (intern pre_config_function) (intern voice_variable)))
129
+ (list 'if (list 'null (intern voice_variable))
130
+ (list 'set! (intern voice_variable)
131
+ (list 'multisyn_load_voice_modules
132
+ (list 'quote name)
133
+ srate
134
+ (list 'quote backoff_rules)
135
+ data_dir
136
+ (list 'quote config))))
137
+ (list (intern config_function) (intern voice_variable))
138
+ (list 'set! 'current-voice (list 'quote name))
139
+ (list 'define_current_voice_reset)
140
+ (list 'set! 'currentMultiSynVoice (intern voice_variable))
141
+ ))
142
+
143
+ (eval (list 'define
144
+ (list (intern free_name))
145
+ (list 'cond
146
+ (list (list 'null (intern voice_variable))
147
+ (list 'error "Voice not currently loaded!"))
148
+ (list (list 'eq? 'currentMultiSynVoice (intern voice_variable))
149
+ (list 'error "Can't free current voice!"))
150
+ (list 't (list set! (intern voice_variable) 'nil))))))
151
+ nil)
152
+
153
+ (define (multisyn_load_voice_modules name srate backoff_rules base_dir module_list)
154
+ "(multisyn_add_module voice name srate backoff_rules base_dir module_list)
155
+ Add voice modules to a voice."
156
+ (let (voice)
157
+ (mapcar
158
+ (lambda (module_entry)
159
+ (let ((dirs (car module_entry))
160
+ (utt_list (load_utt_list (path-append base_dir
161
+ (cadr module_entry)))))
162
+ (if (null voice)
163
+ (set! voice (make_du_voice utt_list dirs srate))
164
+ (voice.addModule voice utt_list dirs srate))))
165
+ module_list)
166
+ (voice.setName voice name)
167
+ (if flattenVoice
168
+ (du_voice.setTargetCost voice "flat")
169
+ (du_voice.setTargetCost voice t))
170
+ (du_voice.setJoinCost voice t)
171
+ (format stderr "Please wait: Initialising multisyn voice.\n")
172
+ (voice.init voice)
173
+ (format stderr " Voice loaded successfully!\n")
174
+ (du_voice.set_ob_pruning_beam voice 0.25)
175
+ (du_voice.set_pruning_beam voice 0.25)
176
+ (du_voice.setDiphoneBackoff voice backoff_rules)
177
+ voice))
178
+
179
+
180
+
181
+
182
+ (define (define_current_voice_reset)
183
+ "(define_current_voice_reset)
184
+ Re-define (current_voice_reset) correctly."
185
+ (eval (list 'define
186
+ (list 'current_voice_reset)
187
+ (list 'multisyn_reset_globals))))
188
+
189
+ (define (multisyn_reset_globals)
190
+ "(multisyn_reset_globals)
191
+ Reset multisyn specific global variables."
192
+ (Param.set 'unisyn.window_symmetric 1))
193
+
194
+
195
+ (provide 'multisyn)
CosyVoice-ttsfrd/resource/festival/multisyn/multisyn_pauses.scm ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 2003, 2004 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Multisyn Pause module (Rob Clark and Korin Richmond)
35
+ ;;;
36
+ ;;;
37
+
38
+ (defvar BB_Pause "B_300")
39
+ (defvar B_Pause "B_150")
40
+ (defvar mB_Pause "B_150") ; shouldn't be used
41
+
42
+ (define (MultiSyn_Pauses utt)
43
+ "(MultiSyn_Pauses UTT)
44
+ Predict pause insertion in a Multisyn unit selection utterance structure."
45
+ (let ((words (utt.relation.items utt 'Word)) lastword tpname)
46
+ (if words
47
+ (begin
48
+ (insert_initial_pause utt) ;; always have a start pause
49
+ (set! lastword (car (last words)))
50
+ (mapcar
51
+ (lambda (w)
52
+ (let ((pbreak (item.feat w "pbreak"))
53
+ (emph (item.feat w "R:Token.parent.EMPH")))
54
+ (cond
55
+ ((string-equal pbreak "BB")
56
+ (unitselection_pause_insert w BB_Pause))
57
+ ((string-equal pbreak "mB")
58
+ (unitselection_pause_insert w mB_Pause))
59
+ ((string-equal pbreak "B")
60
+ (unitselection_pause_insert w B_Pause)))))
61
+ words)
62
+ ;; The embarassing bit. Remove any words labelled as punc or fpunc
63
+ (mapcar
64
+ (lambda (w)
65
+ (let ((pos (item.feat w "pos")))
66
+ (if (or (string-equal "punc" pos)
67
+ (string-equal "fpunc" pos))
68
+ (let ((pbreak (item.feat w "pbreak"))
69
+ (wp (item.relation w 'Phrase)))
70
+ (if (and (string-matches pbreak "BB?")
71
+ (item.relation.prev w 'Word))
72
+ (item.set_feat
73
+ (item.relation.prev w 'Word) "pbreak" pbreak))
74
+ (item.relation.remove w 'Word)
75
+ ;; can't refer to w as we've just deleted it
76
+ (item.relation.remove wp 'Phrase)))))
77
+ words)))
78
+ ;(utt.relation.print utt 'Word)
79
+ ;(utt.relation.print utt 'Segment)
80
+ utt))
81
+
82
+ (define (unitselection_pause_insert word pause)
83
+ "(pause_insert word pause)
84
+ Insert segments needed for a pause."
85
+ (let ((silence (car (cadr (car (PhoneSet.description '(silences))))))
86
+ (seg (item.relation (find_last_seg word) 'Segment))
87
+ pause_item)
88
+ ;(format stderr " inserting pause after: %s.\n" (item.name seg))
89
+ ;(format stderr " Inserting pause\n")
90
+ ; if next seg is not silence insert one.
91
+ (if (or (not (item.next seg))
92
+ (not (string-equal (item.name (item.next seg)) silence)))
93
+ (item.insert seg (list silence) 'after))
94
+ ; insert pause after that if not the end.
95
+ (if (item.next (item.next seg))
96
+ (begin
97
+ (set! pause_item (item.insert (item.next seg) (list pause) 'after))
98
+ ;if next seg after that is not silence add one.
99
+ (if (not (string-equal (item.name (item.next pause_item)) silence))
100
+ (item.insert pause_item (list silence) 'after))))))
101
+
102
+ (provide 'multisyn_pauses)
CosyVoice-ttsfrd/resource/festival/multisyn/radio_phones_multisyn.scm ADDED
@@ -0,0 +1,136 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997,2003, 2004 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; A definition of the radio phone set used in the BU RADIO FM
35
+ ;;; corpus, some people call this the darpa set. This one
36
+ ;;; has the closures removed and pauses added for multisyn
37
+ ;;;
38
+
39
+ (defPhoneSet
40
+ radio_multisyn
41
+ ;;; Phone Features
42
+ (;; vowel or consonant
43
+ (vc + -)
44
+ ;; vowel length: short long dipthong schwa
45
+ (vlng s l d a 0)
46
+ ;; vowel height: high mid low
47
+ (vheight 1 2 3 0)
48
+ ;; vowel frontness: front mid back
49
+ (vfront 1 2 3 0)
50
+ ;; lip rounding
51
+ (vrnd + - 0)
52
+ ;; consonant type: stop fricative affricate nasal lateral approximant
53
+ (ctype s f a n l r 0)
54
+ ;; place of articulation: labial alveolar palatal labio-dental
55
+ ;; dental velar glottal
56
+ (cplace l a p b d v g 0)
57
+ ;; consonant voicing
58
+ (cvox + - 0)
59
+ )
60
+ ;; Phone set members
61
+ (
62
+ ;; multisyn extras
63
+ (# - 0 0 0 0 0 0 -) ;; slience ...
64
+ (B_10 - 0 0 0 0 0 0 -) ;; Pauses
65
+ (B_20 - 0 0 0 0 0 0 -) ;; Pauses
66
+ (B_30 - 0 0 0 0 0 0 -) ;; Pauses
67
+ (B_40 - 0 0 0 0 0 0 -) ;; Pauses
68
+ (B_50 - 0 0 0 0 0 0 -) ;; Pauses
69
+ (B_100 - 0 0 0 0 0 0 -) ;; Pauses
70
+ (B_150 - 0 0 0 0 0 0 -) ;; Pauses
71
+ (B_200 - 0 0 0 0 0 0 -) ;; Pauses
72
+ (B_250 - 0 0 0 0 0 0 -) ;; Pauses
73
+ (B_300 - 0 0 0 0 0 0 -) ;; Pauses
74
+ (B_400 - 0 0 0 0 0 0 -) ;; Pauses
75
+
76
+ ;; Note these features were set by awb so they are wrong !!!
77
+ (aa + l 3 3 - 0 0 0) ;; father
78
+ (ae + s 3 1 - 0 0 0) ;; fat
79
+ (ah + s 2 2 - 0 0 0) ;; but
80
+ (ao + l 3 3 + 0 0 0) ;; lawn
81
+ (aw + d 3 2 - 0 0 0) ;; how
82
+ (ax + a 2 2 - 0 0 0) ;; about
83
+ (axr + a 2 2 - r a +)
84
+ (ay + d 3 2 - 0 0 0) ;; hide
85
+ (b - 0 0 0 0 s l +)
86
+ (ch - 0 0 0 0 a p -)
87
+ (d - 0 0 0 0 s a +)
88
+ (dh - 0 0 0 0 f d +)
89
+ (dx - a 0 0 0 s a +) ;; ??
90
+ (eh + s 2 1 - 0 0 0) ;; get
91
+ (el + s 0 0 0 l a +)
92
+ (em + s 0 0 0 n l +)
93
+ (en + s 0 0 0 n a +)
94
+ (er + a 2 2 - r 0 0) ;; always followed by r (er-r == axr)
95
+ (ey + d 2 1 - 0 0 0) ;; gate
96
+ (f - 0 0 0 0 f b -)
97
+ (g - 0 0 0 0 s v +)
98
+ (hh - 0 0 0 0 f g -)
99
+ (hv - 0 0 0 0 f g +)
100
+ (ih + s 1 1 - 0 0 0) ;; bit
101
+ (iy + l 1 1 - 0 0 0) ;; beet
102
+ (jh - 0 0 0 0 a p +)
103
+ (k - 0 0 0 0 s v -)
104
+ (l - 0 0 0 0 l a +)
105
+ (m - 0 0 0 0 n l +)
106
+ (n - 0 0 0 0 n a +)
107
+ (nx - 0 0 0 0 n d +) ;; ???
108
+ (ng - 0 0 0 0 n v +)
109
+ (ow + d 2 3 + 0 0 0) ;; lone
110
+ (oy + d 2 3 + 0 0 0) ;; toy
111
+ (p - 0 0 0 0 s l -)
112
+ (r - 0 0 0 0 r a +)
113
+ (s - 0 0 0 0 f a -)
114
+ (sh - 0 0 0 0 f p -)
115
+ (t - 0 0 0 0 s a -)
116
+ (th - 0 0 0 0 f d -)
117
+ (uh + s 1 3 + 0 0 0) ;; full
118
+ (uw + l 1 3 + 0 0 0) ;; fool
119
+ (v - 0 0 0 0 f b +)
120
+ (w - 0 0 0 0 r l +)
121
+ (y - 0 0 0 0 r p +)
122
+ (z - 0 0 0 0 f a +)
123
+ (zh - 0 0 0 0 f p +)
124
+ (pau - 0 0 0 0 0 0 -)
125
+ (h# - 0 0 0 0 0 0 -)
126
+ (brth - 0 0 0 0 0 0 -)
127
+ )
128
+ )
129
+
130
+ (PhoneSet.silences '(# pau h# brth))
131
+
132
+ (provide 'radio_phones_multisyn)
133
+
134
+
135
+
136
+
CosyVoice-ttsfrd/resource/festival/multisyn/send_xwaves.scm ADDED
@@ -0,0 +1,318 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 2003, 2004 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; xwaves interface for festival for multisyn (Rob Clark)
35
+ ;;;
36
+ ;;; This is never loaded by defualt.
37
+ ;;; You'd need to change the paths here for this to currently work outside of CSTR.
38
+ ;;; If anyone else ends up using it let me know and I'll make it more robust.
39
+ ;;;
40
+
41
+ ;; Send commands to xwaves
42
+
43
+ (defvar send_xwaves_command "/cstr/linux/entropic/esps531.linux/bin/send_xwaves")
44
+ (defvar spectrogram_command "/cstr/linux/entropic/esps531.linux/bin/sgram")
45
+ (defvar data_path "/projects/cougar/data/cstr/nina")
46
+
47
+ (set! xw_object_count 0)
48
+ (set! xw_active_list nil)
49
+
50
+ ;;
51
+ ;; Display a synthesised utterance
52
+ ;;
53
+ (define (xwaves_display_utterance utt)
54
+ "(xwaves_display_utterance utt)
55
+ Display join and target information for an utterance."
56
+ (let ((units (utt.relation.items utt 'Unit))
57
+ (object (xw_name_object))
58
+ wavfile specfile segfile diphfile joinfile targfile sourcefile timefile)
59
+
60
+ (set! wavfile (xw_make_tmp_filename object))
61
+ (set! specfile (xw_make_tmp_filename object))
62
+ (set! segfile (xw_make_tmp_filename object))
63
+ (set! diphfile (xw_make_tmp_filename object))
64
+ (set! joinfile (xw_make_tmp_filename object))
65
+ (set! targfile (xw_make_tmp_filename object))
66
+ (set! sourcefile (xw_make_tmp_filename object))
67
+ (set! timefile (xw_make_tmp_filename object))
68
+
69
+ ; display resulting waveform
70
+ (utt.save.wave utt wavfile 'riff)
71
+ (xwaves_show_general object wavfile 1500 200 10 10)
72
+ ; display resulting spectrogram
73
+ (xw_genspec wavfile specfile)
74
+ (xwaves_show_general object specfile 1500 400 10 260)
75
+ ; segments
76
+ (utt.save.unit_selection_segs utt segfile)
77
+ (xwaves_show_labels object segfile specfile)
78
+ ; Unit information
79
+ (utt.save.unit_selection_info utt diphfile joinfile targfile sourcefile timefile)
80
+ (xwaves_show_labels object timefile specfile)
81
+ (xwaves_show_labels object sourcefile specfile)
82
+ (xwaves_show_labels object targfile specfile)
83
+ (xwaves_show_labels object joinfile specfile)
84
+ (xwaves_show_labels object diphfile specfile)
85
+ ; mark files
86
+ (xw_register_active object (list wavfile specfile segfile diphfile joinfile sourcefile timefile))
87
+ ))
88
+
89
+ ;;
90
+ ;; Edit a diphone source
91
+ ;;
92
+
93
+ (define (xwaves_edit_diphone utt id)
94
+ "(xwaves_edit_diphone utt id)
95
+ Access the source diphone for label correction."
96
+ (let ((diphone nil)
97
+ segfilename
98
+ wavefilename
99
+ (utt (Utterance Text nil))
100
+ segs
101
+ (seg nil)
102
+ (start 0)
103
+ end)
104
+
105
+ ;; find unit.
106
+ (mapcar
107
+ (lambda (unit)
108
+ (if (string-equal (format nil "_%s" id) (item.feat unit id))
109
+ (set! diphone unit)))
110
+ (utt.relation.items utt 'Unit))
111
+ (if (null diphone)
112
+ (error (format nil "Diphone with id _%s not found in utterance.")))
113
+ (set! uttname (item.feat diphone "source_utt"))
114
+ (set! end (item.feat diphone "source_end"))
115
+
116
+ (set! segfilename (format nil "%s/lab/%s.lab" data_path uttname))
117
+ (set! wavefilename (format nil "%s/wav/%s.wav" data_path uttname))
118
+ (utt.relation.load utt 'Segment segfilename)
119
+
120
+ (set! segs (utt.relation.items utt 'Segment))
121
+ (while (and segs
122
+ (not (equal? (item.feat (car segs) "end") end)))
123
+ (set! segs (cdr segs)))
124
+
125
+ if null seg ...
126
+
127
+ (if (item.prev diphone)
128
+ (set! start (item.feat seg "start"))
129
+ (set! start 0))
130
+
131
+
132
+ ))
133
+
134
+
135
+
136
+
137
+
138
+ ;;
139
+ ;; Interface with xwaves.
140
+ ;;
141
+
142
+
143
+ (define (xwaves_show_general object file width height xpos ypos)
144
+ "(xwaves_show_general object file width height xpos ypos)
145
+ Display an wave or track file."
146
+ (xw_send (format nil "make name %s file %s width %d height %d loc_x %d loc_y %d" object file width height xpos ypos)))
147
+
148
+ (define (xwaves_show_wave object file)
149
+ "(xwaves_show_wave object file)
150
+ Display a waveform."
151
+ (xwaves_show_general object file 1500 200 10 10))
152
+
153
+ (define (xwaves_show_labels object file attachto)
154
+ "(xwaves_show_labels object file attachto)
155
+ Display a label file."
156
+ (xw_send (format nil "send make signal %s name %s file %s color 125" attachto object file))
157
+ (xw_send "send activate fields 1 2 3 4 5"))
158
+
159
+
160
+ (define (xwaves_attach_xlabel)
161
+ "(xwaves_attach_xlabel)
162
+ Attach xlabel to xwaves."
163
+ (xw_send "attach function xlabel"))
164
+
165
+ (define (xwaves_set_markers object left right)
166
+ "(xwaves_set_markers object left right)
167
+ Set the markers."
168
+ (xw_send (format nil "%s set l_marker_time %f" object left))
169
+ (xw_send (format nil "%s set r_marker_time %f" object right)))
170
+
171
+ (define (xwaves_bracket_markers object file)
172
+ "(xwaves_bracket_markers object file)
173
+ Bracket markers."
174
+ (xw_send (format nil "%s bracket file %s " object file)))
175
+
176
+ (define (xwaves_close_windows object)
177
+ "(xwaves_close_windows object)
178
+ Close currently open windows related to object or all if nil.."
179
+ (cond
180
+ ((null object)
181
+ (xw_send "kill"))
182
+ (t
183
+ (xw_send (format nil "kill name %s" object))))
184
+ (xw_clear_active_list object))
185
+
186
+
187
+ (define (xwaves_wait)
188
+ "(xwaves_wait)
189
+ Wait for xwaves continue signal."
190
+ (xw_send "pause"))
191
+
192
+
193
+ ;;
194
+ ;; Object naming
195
+ ;;
196
+ (define (xw_name_object)
197
+ "(xw_name_object)
198
+ Generate a name for this object."
199
+ (let (name)
200
+ (set! name (string-append "obj" xw_object_count))
201
+ (set! xw_object_count (+ xw_object_count 1))
202
+ name))
203
+
204
+ ;;
205
+ ;; Temp file lists
206
+ ;;
207
+
208
+ (define (xw_clear_active_list object)
209
+ "(xw_clear_active_list)
210
+ Clear active list of specified object, or all if nil."
211
+ (let (new_active_list)
212
+ (mapcar
213
+ (lambda (objectlist)
214
+ (cond
215
+ ((or (null object)
216
+ (string-equal object (car objectlist)))
217
+ (mapcar
218
+ (lambda (file)
219
+ (delete-file file))
220
+ (cadr objectlist)))
221
+ (t
222
+ (set! new_active_list (cons objectlist new_active_list)))))
223
+ xw_active_list)
224
+ (set! xw_active_list new_active_list))
225
+ nil)
226
+
227
+
228
+ (define (xw_register_active object flist)
229
+ "(xw_register_active object flist)
230
+ Adds an object and its filenames to the active list."
231
+ (set! xw_active_list (cons (cons object (list flist)) xw_active_list))
232
+ nil)
233
+
234
+ (define (xw_make_tmp_filename object)
235
+ "(xw_make_tmp_filename)
236
+ make tmp file name which incorporates object name."
237
+ (format nil "%s_%s" (make_tmp_filename) object))
238
+
239
+
240
+ ;;
241
+ ;; Low level xwaves stuff.
242
+ ;;
243
+
244
+ (define (xw_genspec wavfile specfile)
245
+ "(xw_genspec wavfile specfile)
246
+ Generate a spectrogram file."
247
+ (system (format nil "%s -dHAMMING -o8 -E0.94 -S2 -w8 %s %s\n" spectrogram_command wavfile specfile)))
248
+
249
+ (define (xw_send command)
250
+ "(xw_send command)
251
+ Send a command to xwaves."
252
+ (system (format nil "%s %s\n" send_xwaves_command command)))
253
+
254
+
255
+
256
+ ;;
257
+ ;; General Festival stuff.
258
+ ;;
259
+
260
+
261
+ (define (utt.save.unit_selection_segs utt filename)
262
+ "(utt.save.unit_selection_segs utt filename)
263
+ Save unit selection segments of UTT in a FILE in xlabel format."
264
+ (let ((fd (fopen filename "w")))
265
+ (format fd "#\n")
266
+ (mapcar
267
+ (lambda (info)
268
+ (format fd "%2.4f 100 %s\n" (car info) (car (cdr info))))
269
+ (utt.features utt 'Segment '(source_end name)))
270
+ (fclose fd)
271
+ utt))
272
+
273
+ (define (utt.save.unit_selection_info utt diphfile joinfile targfile sourcefile timefile)
274
+ "(utt.save.unit_selection_info utt diphfile joinfile targfile sourcefile timefile)
275
+ Save stuff in xlabel format."
276
+ (let ((fdd (fopen diphfile "w"))
277
+ (fdj (fopen joinfile "w"))
278
+ (fdt (fopen targfile "w"))
279
+ (fds (fopen sourcefile "w"))
280
+ (fdx (fopen timefile "w"))
281
+ real_join)
282
+ (format fdd "#\n")
283
+ (format fdj "#\n")
284
+ (format fdt "#\n")
285
+ (format fds "#\n")
286
+ (format fdx "#\n")
287
+ (mapcar
288
+ (lambda (unit)
289
+ (set! real_join "")
290
+ (if (item.next unit)
291
+ (if (not (string-equal (item.feat unit 'source_utt)
292
+ (item.feat (item.next unit) 'source_utt)))
293
+ (set! real_join "*")))
294
+ (format fdd "%2.4f 100 %s %s\n"
295
+ (item.feat unit 'end)
296
+ (item.feat unit 'name)
297
+ real_join)
298
+ (format fdj "%2.4f 100 %s\n"
299
+ (item.feat unit 'end)
300
+ (if (item.next unit)
301
+ (item.feat (item.next unit) 'join_cost)
302
+ 0))
303
+ (format fdt "%2.4f 100 %s\n"
304
+ (item.feat unit 'end)
305
+ (item.feat unit 'target_cost))
306
+ (format fds "%2.4f 100 %s\n"
307
+ (item.feat unit 'end)
308
+ (item.feat unit 'source_utt))
309
+ (format fdx "%2.4f 100 %s\n"
310
+ (item.feat unit 'end)
311
+ (item.feat unit 'source_end)))
312
+ (utt.relation.items utt 'Unit))
313
+ (fclose fdd)
314
+ (fclose fdj)
315
+ (fclose fdt)
316
+ (fclose fds)
317
+ (fclose fdx)
318
+ utt))
CosyVoice-ttsfrd/resource/festival/multisyn/target_cost.scm ADDED
@@ -0,0 +1,410 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 2003, 2004 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Multisyn scheme target cost (Rob Clark and Korin Richmond)
35
+ ;;;
36
+ ;;;
37
+
38
+ (define (Default_Target_Cost targ cand)
39
+ "(Default_Target_Cost targ cand)
40
+ A Default Target Cost function."
41
+ (let ((cost 0))
42
+ (mapcar
43
+ (lambda (row)
44
+ (set! cost (+ cost (tc_eval_row row targ cand))))
45
+ target_matrix)
46
+ (set! cost (/ cost target_matrix_weight))
47
+ cost))
48
+
49
+
50
+ (define (tc_eval_row row targ cand)
51
+ "(tc_eval_row row targ cand)
52
+ Evaluate a target matrix row."
53
+ (let ((weight (car row))
54
+ (func (cadr row))
55
+ (result 0))
56
+ (set! result (* weight (eval (list func targ cand))))
57
+ result))
58
+
59
+ ;;
60
+ ;; Target cost Matrix
61
+ ;; '(weight function)
62
+
63
+ (define (get_matrix_weight m)
64
+ (let ((w 0))
65
+ (mapcar
66
+ (lambda (x)
67
+ (set! w (+ w (car x))))
68
+ m)
69
+ w))
70
+
71
+
72
+ (set! test_matrix_max_weight 1)
73
+ (set! test_matrix
74
+ '(
75
+ (10 tc_stress )
76
+ (5 tc_syl_pos )
77
+ (5 tc_word_pos)
78
+ (6 tc_partofspeech)
79
+ (7 tc_phrase_pos)
80
+ (4 tc_left_context)
81
+ (3 tc_right_context)
82
+ (25 tc_bad_f0) ;; set to equal 1/3 of total cost (so high because interaction with join)
83
+ ; (0 tc_segment_score) ;; was 4. turned off until utterances are built for this.
84
+ (10 tc_bad_duration) ;; was 6
85
+ ))
86
+
87
+ (set! test_matrix_weight (* test_matrix_max_weight (get_matrix_weight test_matrix)))
88
+
89
+ (set! target_matrix test_matrix)
90
+ (set! target_matrix_weight test_matrix_weight)
91
+
92
+
93
+
94
+ ;;
95
+ ;; tc_stress
96
+ ;;
97
+ ;; Compares stress on any vowel which form part of the diphone. stress
98
+ ;; conditions must match for a zero target cost.
99
+ ;;
100
+
101
+ (define (tc_stress targ cand)
102
+ "(tc_stress targ cand)
103
+ Target Cost stressed. 0 - stress patterns match [ compares: 0 unstressed vs. > 0 stressed ]
104
+ 1 - stress miss-match.
105
+ "
106
+ (let ((c 0)
107
+ cand_stress targ_stress)
108
+ ;(format t "my_is_vowel %l\n" (my_is_vowel targ))
109
+ ;(format t "phone_is_silence %l\n" (phone_is_silence (item.feat targ 'name)))
110
+ ;; For first segment
111
+ (if (and (not (phone_is_silence (item.feat targ 'name)))
112
+ (my_is_vowel targ))
113
+ (begin
114
+ (set! cand_stress (item.feat cand "R:SylStructure.parent.stress"))
115
+ (set! targ_stress (item.feat targ "R:SylStructure.parent.stress"))
116
+ (if (or (and (eq? cand_stress 0) (> targ_stress 0))
117
+ (and (eq? targ_stress 0) (> cand_stress 0)))
118
+ (set! c 1))))
119
+ ;; For second segment
120
+ ;(format t "n.my_is_vowel %l\n" (my_is_vowel (item.next targ)))
121
+ ;(format t "n.phone_is_silence %l\n" (phone_is_silence (item.feat targ 'n.name)))
122
+ (if (and (not (phone_is_silence (item.feat targ 'n.name)))
123
+ (my_is_vowel (item.next targ)))
124
+ (begin
125
+ (set! cand_stress (item.feat cand "n.R:SylStructure.parent.stress"))
126
+ (set! targ_stress (item.feat targ "n.R:SylStructure.parent.stress"))
127
+ (if (or (and (eq? cand_stress 0) (> targ_stress 0))
128
+ (and (eq? targ_stress 0) (> cand_stress 0)))
129
+ (set! c 1))))
130
+ ; (format t "tc_stress: %l\n" c)
131
+ c))
132
+
133
+
134
+ ;;
135
+ ;; tc_syl_position
136
+ ;;
137
+ ;; Find and compare diphone position in syllabic structure.
138
+ ;; Values are: inter - diphone crosses syllable boundary.
139
+ ;; initial - diphone is syllable initial.
140
+ ;; medial - diphone is syllable medial
141
+ ;; final - diphone is syllable final
142
+ ;; returns 0 for a match 1 for a mismatch.
143
+ ;;
144
+ (define (tc_syl_pos targ cand)
145
+ "(tc_syl_pos targ cand)
146
+ Score position in syllable."
147
+ (let ((targ_pos "medial")
148
+ (cand_pos "medial")
149
+ (targ_syl (get_syl targ))
150
+ (targ_next_syl (get_syl (item.next targ)))
151
+ (cand_syl (get_syl cand))
152
+ (cand_next_syl (get_syl (item.next cand))))
153
+ ;; target
154
+ (cond
155
+ ((not (equal? targ_syl targ_next_syl))
156
+ (set! targ_pos "inter"))
157
+ ((not (equal? targ_syl (get_syl (item.prev targ))))
158
+ (set! targ_pos "initial"))
159
+ ((not (equal? targ_next_syl (get_syl (item.next (item.next targ)))))
160
+ (set! targ_pos "final")))
161
+ ;; candidate
162
+ (cond
163
+ ((not (equal? cand_syl cand_next_syl))
164
+ (set! cand_pos "inter"))
165
+ ((not (equal? cand_syl (get_syl (item.prev cand))))
166
+ (set! cand_pos "initial"))
167
+ ((not (equal? cand_next_syl (get_syl (item.next (item.next cand)))))
168
+ (set! cand_pos "final")))
169
+ ; (format t "targ_syl: %l cand_syl %l\n" targ_pos cand_pos)
170
+ (if (equal? targ_pos cand_pos) 0 1)))
171
+
172
+ ;;
173
+ ;; tc_word_position
174
+ ;;
175
+ ;; Find and compare diphone position in word structure
176
+ ;; Values are: inter - diphone crosses word boundary.
177
+ ;; initial - diphone is word initial.
178
+ ;; medial - diphone is word medial
179
+ ;; final - diphone is word final
180
+ ;; returns 0 for a match 1 for a mismatch.
181
+ ;;
182
+ (define (tc_word_pos targ cand)
183
+ "(tc_word_pos targ cand)
184
+ Score position in word."
185
+ (let ((targ_pos "medial")
186
+ (cand_pos "medial")
187
+ (targ_word (get_word targ))
188
+ (targ_next_word (get_word (item.next targ)))
189
+ (cand_word (get_word cand))
190
+ (cand_next_word (get_word (item.next cand))))
191
+ ;; target
192
+ (cond
193
+ ((not (equal? targ_word targ_next_word))
194
+ (set! targ_pos "inter"))
195
+ ((not (equal? targ_word (get_word (item.prev targ))))
196
+ (set! targ_pos "initial"))
197
+ ((not (equal? targ_next_word (get_word (item.next (item.next targ)))))
198
+ (set! targ_pos "final")))
199
+ ;; candidate
200
+ (cond
201
+ ((not (equal? cand_word cand_next_word))
202
+ (set! cand_pos "inter"))
203
+ ((not (equal? cand_word (get_word (item.prev cand))))
204
+ (set! cand_pos "initial"))
205
+ ((not (equal? cand_next_word (get_word (item.next (item.next cand)))))
206
+ (set! cand_pos "final")))
207
+ ; (format t "targ_word: %l cand_word %l\n" targ_pos cand_pos)
208
+ (if (equal? targ_pos cand_pos) 0 1)))
209
+
210
+
211
+
212
+ ;;
213
+ ;; tc_phrase_position
214
+ ;;
215
+ ;; Position (of word) in phrase
216
+ ;; initial/medial/final
217
+ ;;
218
+ ;; 0 - match, 1 - mismatch
219
+ ;;
220
+ (define (tc_phrase_pos targ cand)
221
+ "(tc_phrase_pos targ cand)
222
+ Score position in phrase."
223
+ (let ((targ_word (get_word targ))
224
+ (cand_word (get_word cand)))
225
+ (cond
226
+ ((and (null targ_word)
227
+ (null cand_word))
228
+ 0)
229
+ ((or (null targ_word)
230
+ (null cand_word))
231
+ 1)
232
+ ((string-equal (item.feat targ_word 'pbreak)
233
+ (item.feat cand_word 'pbreak))
234
+ 0)
235
+ (t 1))))
236
+
237
+ ;;
238
+ ;; tc_partofspeech
239
+ ;;
240
+ ;;
241
+ ;;
242
+ (define (tc_partofspeech targ cand)
243
+ "(tc_partofspeech targ cand)
244
+ Score part of speech."
245
+ (let ((targ_word (get_word targ))
246
+ (cand_word (get_word cand))
247
+ targ_pos cand_pos)
248
+ (if targ_word
249
+ (set! targ_pos (simple_pos (item.feat targ_word 'pos))))
250
+ (if cand_word
251
+ (set! cand_pos (simple_pos (item.feat cand_word 'pos))))
252
+ ;(format t "targ_pos %l cand_pos %l\n" targ_pos cand_pos)
253
+ (if (equal? targ_pos cand_pos) 0 1)))
254
+
255
+ (define (score_contexts targ_context cand_context)
256
+ "(score_contexts targ_context cand_context)
257
+ If both context items are nil, then score is 0.
258
+ If both context items are not nil, and are the same, then
259
+ score is 0. Otherwise, score is 1."
260
+ (if (and targ_context cand_context)
261
+ (if (equal? (item.feat targ_context "name")
262
+ (item.feat cand_context "name"))
263
+ 0
264
+ 1)
265
+ (if (and (equal? targ_context nil)
266
+ (equal? cand_context nil))
267
+ 0
268
+ 1)))
269
+
270
+
271
+ (define (tc_left_context targ cand)
272
+ "(tc_left_context targ cand)
273
+ Score left phonetic context."
274
+ (let ((targ_context (item.prev targ))
275
+ (cand_context (item.prev cand)))
276
+ (score_contexts targ_context cand_context)))
277
+
278
+ ;;
279
+ ;; tc_right_context
280
+ ;;
281
+ ;;
282
+ ;;
283
+ (define (tc_right_context targ cand)
284
+ "(tc_right_context targ cand)
285
+ Score right phonetic context."
286
+ (let ((targ_context (item.next (item.next targ)))
287
+ (cand_context (item.next (item.next cand))))
288
+ (score_contexts targ_context cand_context)))
289
+
290
+
291
+ ;;
292
+ ;; tc_segment_score
293
+ ;;
294
+ ;; This currently thresholds based on looking at the distributions of the scores.
295
+ ;; A nice exp function may be better.
296
+ (define (tc_segment_score targ cand)
297
+ "tc_segment_score targ cand)
298
+ A bad alignment score make a bad segment."
299
+ (let ((score 0))
300
+ (if (not (phone_is_silence (item.feat cand "name")))
301
+ (set! score (+ score (item.feat cand 'score))))
302
+ (if (not (phone_is_silence (item.feat (item.next cand) "name")))
303
+ (set! score (+ score (item.feat (item.next cand) 'score))))
304
+ (cond
305
+ ((> score -4000) ;2000 (x2) is 7.5%
306
+ 0)
307
+ ((> score -5000) ;2500 (x2) is 5.0%
308
+ 0.5)
309
+ (t 1))))
310
+
311
+ ;;
312
+ ;; tc_bad_duration
313
+ ;;
314
+ ;; If the segment is marked as having a weird duration penalise it.
315
+ ;; We allow bad_dur to be set on the target so resynthesis works
316
+ ;; and so you could ask for really long/short segments.
317
+ ;;
318
+ (define (tc_bad_duration targ cand)
319
+ (if (equal? (item.feat targ "bad_dur")
320
+ (item.feat cand "bad_dur"))
321
+ 0
322
+ 1))
323
+
324
+
325
+ ;;
326
+ ;; tc_bad_f0
327
+ ;;
328
+ ;; If the candidate is deemed to have an inappropriate f0, then penalise it.
329
+ ;;
330
+ ;; Specifically, if the targ/cand segment type is expected to be voiced, then
331
+ ;; an f0 of zero is bad (results from poor pitch tracking). In such a case,
332
+ ;; the join cost would then favour other units with f0 (since the euclidean
333
+ ;; distance between two zeros is very small ;)
334
+ ;; We want to avoid that.
335
+ ;;
336
+ ;; Presumeably, we also want to penalise cases where supposedly voiceless
337
+ ;; candidates have an f0 != 0 (either a consequence of bad pitch tracking
338
+ ;; or bad labelling) but that's not done here yet...
339
+ ;;
340
+ ;; (the function itself has been implemented in C for convenience, and
341
+ ;; this stub is left here just for this note ;)
342
+
343
+ (define (tc_bad_f0 targ cand)
344
+ (let ((score (temp_tc_bad_f0 targ cand))
345
+ (name (format nil "%s_%s"
346
+ (item.feat targ "name")
347
+ (item.feat (item.next targ) "name"))))
348
+ (if (not (equal? score 0.0))
349
+ (format t "f0 score for %s is %f\n" name score))
350
+ score))
351
+
352
+ ;;
353
+ ;; Is a segment a vowel? ( ph_is_a_vowel doesn't seem to work)
354
+ ;;
355
+ (define (my_is_vowel seg)
356
+ (if seg
357
+ (if (equal? (item.feat seg 'ph_vc) "+")
358
+ t
359
+ nil)))
360
+
361
+
362
+
363
+ ;; get the syllable from sysstructure in normal utterance
364
+ ;;
365
+ (define (get_syl seg)
366
+ (let (syl)
367
+ (if seg
368
+ (set! syl (item.relation.parent seg 'SylStructure)))
369
+ syl))
370
+
371
+ ;; get the word from sylstructure in normal utterance
372
+ ;;
373
+ (define (get_word seg)
374
+ (let ((syl (get_syl seg))
375
+ word)
376
+ (if syl
377
+ (set! word (item.parent syl)))
378
+ word))
379
+
380
+
381
+ ;; simple pos
382
+ ;;
383
+ (define (simple_pos pos)
384
+ (let (spos)
385
+ (cond
386
+ ((member_string pos '(vbd vb vbn vbz vbp vbg))
387
+ (set! spos "v"))
388
+ ((member_string pos '(nn nnp nns nnps fw sym ls))
389
+ (set! spos "n"))
390
+ ((member_string pos '(dt gin prp cc of to cd md pos wdt wp wrb ex uh pdt))
391
+ (set! spos "func"))
392
+ ((member_string pos '(jj jjr jjs 1 2 rb rp rbr rbs))
393
+ (set! spos "other")))
394
+ spos))
395
+
396
+
397
+ ;; debugging
398
+
399
+ (define (test_target_cost utt1 utt2)
400
+ (let ((segs1 (utt.relation.items utt1 'Segment))
401
+ (segs2 (utt.relation.items utt2 'Segment))
402
+ (tc 0))
403
+ (while (and segs1 segs2)
404
+ (set! tc (Default_Target_Cost (car segs1) (car segs2)))
405
+ (format t "targ: %l cand: %l cost: %l\n" (item.name (car segs1)) (item.name (car segs2)) tc)
406
+ (set! segs1 (cdr segs1))
407
+ (set! segs2 (cdr segs2)))))
408
+
409
+
410
+ (provide 'target_cost)
CosyVoice-ttsfrd/resource/festival/sable-mode.scm ADDED
@@ -0,0 +1,560 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1998 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;; ;;
34
+ ;;; Festival (1.3.X) support for SABLE 0.2 the SGML/XML based mark up ;;
35
+ ;;; language. ;;
36
+ ;;; ;;
37
+ ;;; This is XML version requiring Edinburgh's LTG's rxp XML parser as ;;
38
+ ;;; distributed with Festival ;;
39
+ ;;; ;;
40
+
41
+ (require_module 'rxp)
42
+
43
+ ;;(set! auto-text-mode-alist
44
+ ;; (cons
45
+ ;; (cons "\\.sable$" 'sable)
46
+ ;; auto-text-mode-alist))
47
+
48
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
49
+ ;; ;;
50
+ ;; Remember where to find these two XML entities. ;;
51
+ ;; ;;
52
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
53
+
54
+
55
+ (xml_register_id "-//SABLE//DTD SABLE speech mark up//EN"
56
+ (path-append libdir "Sable.v0_2.dtd")
57
+ )
58
+
59
+ (xml_register_id "-//SABLE//ENTITIES Added Latin 1 for SABLE//EN"
60
+ (path-append libdir "sable-latin.ent")
61
+ )
62
+
63
+ ;; (print (xml_registered_ids))
64
+
65
+ (defvar SABLE_RXDOUBLE "-?\\(\\([0-9]+\\.[0-9]*\\)\\|\\([0-9]+\\)\\|\\(\\.[0-9]+\\)\\)\\([eE][---+]?[0-9]+\\)?")
66
+
67
+ (defvar sable_pitch_base_map
68
+ '((highest 1.2)
69
+ (high 1.1)
70
+ (medium 1.0)
71
+ (default 1.0)
72
+ (low 0.9)
73
+ (lowest 0.8)))
74
+ (defvar sable_pitch_med_map
75
+ '((highest 1.2)
76
+ (high 1.1)
77
+ (medium 1.0)
78
+ (default 1.0)
79
+ (low 0.9)
80
+ (lowest 0.8)))
81
+ (defvar sable_pitch_range_map
82
+ '((largest 1.2)
83
+ (large 1.1)
84
+ (medium 1.0)
85
+ (default 1.0)
86
+ (small 0.9)
87
+ (smallest 0.8)))
88
+ (defvar sable_rate_speed_map
89
+ '((fastest 1.5)
90
+ (fast 1.2)
91
+ (medium 1.0)
92
+ (default 1.0)
93
+ (slow 0.8)
94
+ (slowest 0.6)))
95
+ (defvar sable_volume_level_map
96
+ '((loudest 2.0)
97
+ (loud 1.5)
98
+ (default 1.0)
99
+ (medium 1.0)
100
+ (quiet 0.5)))
101
+
102
+ (define (sable_init_globals)
103
+ (set! utts nil)
104
+ (set! sable_omitted_mode nil)
105
+ (set! sable_word_features_stack nil)
106
+ (set! sable_pitch_context nil)
107
+ (set! sable_vol_context nil)
108
+ (set! sable_vol_type 'no_change)
109
+ (set! sable_vol_factor 1.0)
110
+ (set! sable_current_language 'britishenglish)
111
+ (set! sable_unsupported_language nil)
112
+ (set! sable_language_stack nil)
113
+ (set! sable_current_speaker 'voice_kal_diphone)
114
+ (set! sable_speaker_stack nil)
115
+ )
116
+
117
+ (define (sable_token_to_words token name)
118
+ "(sable_token_to_words utt token name)
119
+ SABLE mode token specific analysis."
120
+ (cond
121
+ ((or sable_omitted_mode sable_unsupported_language)
122
+ ;; don't say anything (whole utterance)
123
+ nil)
124
+ ((string-equal "1" (item.feat token "done_sable_sub"))
125
+ ;; to catch recursive calls this when splitting up sub expressions
126
+ (sable_previous_token_to_words token name))
127
+ ((and (not (string-equal "0" (item.feat token "sable_sub")))
128
+ (string-equal "0" (item.feat token "p.sable_sub")))
129
+ (let (words (sub (item.feat token "sable_sub")))
130
+ (item.set_feat token "done_sable_sub" "1")
131
+ (set! words
132
+ (apply append
133
+ (mapcar
134
+ (lambda (w)
135
+ (set! www (sable_previous_token_to_words token w))
136
+ www)
137
+ (read-from-string sub))))
138
+ (item.set_feat token "done_sable_sub" "0")
139
+ words))
140
+ ((string-equal "1" (item.feat token "sable_ignore"))
141
+ ;; don't say anything (individual word)
142
+ nil)
143
+ ((string-equal "1" (item.feat token "sable_ipa"))
144
+ ;; Each token is an IPA phone
145
+ (item.set_feat token "phonemes" (sable-map-ipa name))
146
+ (list name))
147
+ ((string-equal "1" (item.feat token "sable_literal"))
148
+ ;; Only deal with spell here
149
+ (let ((subwords) (subword))
150
+ (item.set_feat token "pos" token.letter_pos)
151
+ (mapcar
152
+ (lambda (letter)
153
+ ;; might be symbols or digits
154
+ (set! subword (sable_previous_token_to_words token letter))
155
+ (if subwords
156
+ (set! subwords (append subwords subword))
157
+ (set! subwords subword)))
158
+ (symbolexplode name))
159
+ subwords))
160
+ ((not (string-equal "0" (item.feat token "token_pos")))
161
+ ;; bypass the prediction stage, if English
162
+ (if (member_string (Parameter.get 'Language)
163
+ '(britishenglish americanenglish))
164
+ (builtin_english_token_to_words token name)
165
+ (sable_previous_token_to_words token name)))
166
+ ;; could be others here later
167
+ (t
168
+ (sable_previous_token_to_words token name))))
169
+
170
+ (defvar sable_elements
171
+ '(
172
+ ("(SABLE" (ATTLIST UTT)
173
+ (eval (list sable_current_speaker)) ;; so we know what state we start in
174
+ (sable_setup_voice_params)
175
+ nil
176
+ )
177
+ (")SABLE" (ATTLIST UTT)
178
+ (xxml_synth UTT) ;; Synthesis the remaining tokens
179
+ nil
180
+ )
181
+ ;; Utterance break elements
182
+ ("(LANGUAGE" (ATTLIST UTT)
183
+ ;; Status: probably complete
184
+ (xxml_synth UTT)
185
+ (set! sable_language_stack
186
+ (cons
187
+ (list sable_current_language sable_unsupported_language)
188
+ sable_language_stack))
189
+ ;; Select a new language
190
+ (let ((language (upcase (car (xxml_attval "ID" ATTLIST)))))
191
+ (cond
192
+ ((or (string-equal language "SPANISH")
193
+ (string-equal language "ES"))
194
+ (set! sable_current_language 'spanish)
195
+ (set! sable_unsupported_language nil)
196
+ (select_language 'spanish))
197
+ ((or (string-equal language "ENGLISH")
198
+ (string-equal language "EN"))
199
+ (set! sable_current_language 'britishenglish)
200
+ (set! sable_unsupported_language nil)
201
+ (select_language 'britishenglish))
202
+ (t ;; skip languages you don't know
203
+ ;; BUG: if current language isn't English this wont work
204
+ (apply_hooks tts_hooks
205
+ (eval (list 'Utterance 'Text
206
+ (string-append "Some text in " language))))
207
+ (set! sable_unsupported_language t)))
208
+ nil))
209
+ (")LANGUAGE" (ATTLIST UTT)
210
+ (xxml_synth UTT)
211
+ (set! sable_unsupported_language (car (cdr (car sable_language_stack))))
212
+ (set! sable_current_language (car (car sable_language_stack)))
213
+ (set! sable_language_stack (cdr sable_language_stack))
214
+ (if (not sable_omitted_mode)
215
+ (begin
216
+ (select_language sable_current_language)
217
+ (sable_setup_voice_params)))
218
+ nil)
219
+ ("(SPEAKER" (ATTLIST UTT)
220
+ ;; Status: GENDER/AGE ignored, should be done by sable-def-speaker
221
+ ;; function to define Festival voices to SABLE
222
+ (xxml_synth UTT)
223
+ (set! sable_speaker_stack (cons sable_current_speaker sable_speaker_stack))
224
+ (cond
225
+ ((not equal? sable_current_language 'britishenglish)
226
+ (print "SABLE: choosen unknown voice, current voice unchanged"))
227
+ ((equal? (car (xxml_attval "NAME" ATTLIST)) 'male1)
228
+ (set! sable_current_speaker 'voice_kal_diphone)
229
+ (voice_kal_diphone))
230
+ ((equal? (car (xxml_attval "NAME" ATTLIST)) 'male2)
231
+ (set! sable_current_speaker 'voice_cmu_us_rms_cg)
232
+ (voice_cmu_us_rms_cg))
233
+ ((equal? (car (xxml_attval "NAME" ATTLIST)) 'male3)
234
+ (set! sable_current_speaker 'voice_ked_diphone)
235
+ (voice_ked_diphone))
236
+ ((equal? (car (xxml_attval "NAME" ATTLIST)) 'male4)
237
+ (set! sable_current_speaker 'voice_rab_diphone)
238
+ (voice_rab_diphone))
239
+ ((equal? (car (xxml_attval "NAME" ATTLIST)) 'male5)
240
+ (set! sable_current_speaker 'voice_cmu_us_awb_cg)
241
+ (voice_cmu_us_awb_cg))
242
+ ((equal? (car (xxml_attval "NAME" ATTLIST)) 'female1)
243
+ (set! sable_current_speaker 'voice_cmu_us_slt_cg)
244
+ (voice_us1_mbrola))
245
+ (t
246
+ (set! sable_current_speaker (intern (string-append "voice_" (car (xxml_attval "NAME" ATTLIST)))))
247
+ (eval (list sable_current_speaker))))
248
+ (sable_setup_voice_params)
249
+ nil)
250
+ (")SPEAKER" (ATTLIST UTT)
251
+ (xxml_synth UTT)
252
+ (set! sable_utt UTT)
253
+ (set! sable_current_speaker (car sable_speaker_stack))
254
+ (set! sable_speaker_stack (cdr sable_speaker_stack))
255
+ (eval (list sable_current_speaker))
256
+ (sable_setup_voice_params)
257
+ nil)
258
+ ("BREAK" (ATTLIST UTT)
259
+ ;; Status: probably complete
260
+ ;; may cause an utterance break
261
+ (let ((level (upcase (car (xxml_attval "LEVEL" ATTLIST)))))
262
+ (cond
263
+ ((null UTT) nil)
264
+ ((string-equal "LARGE" level)
265
+ (xxml_synth UTT)
266
+ nil)
267
+ (t
268
+ (let ((last_token (utt.relation.last UTT'Token)))
269
+ (if last_token
270
+ (item.set_feat last_token "pbreak" "B"))
271
+ UTT)))))
272
+ ("(DIV" (ATLIST UTT)
273
+ ;; Status: probably complete
274
+ (xxml_synth UTT)
275
+ nil)
276
+ ("AUDIO" (ATTLIST UTT)
277
+ ;; Status: MODE (background) ignored, only insertion supported
278
+ ;; mime type of file also ignored, as its LEVEL
279
+ (let ((tmpfile (make_tmp_filename)))
280
+ ;; ignoring mode-background (and will for sometime)
281
+ ;; ignoring level option
282
+ (xxml_synth UTT) ;; synthesizing anything ready to be synthesized
283
+ (get_url (car (xxml_attval "SRC" ATTLIST)) tmpfile)
284
+ (apply_hooks tts_hooks
285
+ (eval (list 'Utterance 'Wave tmpfile)))
286
+ (delete-file tmpfile)
287
+ nil))
288
+ ("(EMPH" (ATTLIST UTT)
289
+ ;; Status: nesting makes no difference, levels ignored
290
+ ;; Festival is particularly bad at adding specific emphasis
291
+ ;; that's what happens when you use statistical methods that
292
+ ;; don't include any notion of emphasis
293
+ ;; This is *not* recursive and only one level of EMPH supported
294
+ (sable_push_word_features)
295
+ (set! xxml_word_features
296
+ (cons (list "dur_stretch" 1.6)
297
+ (cons
298
+ (list "EMPH" "1") xxml_word_features)))
299
+ UTT)
300
+ (")EMPH" (ATTLIST UTT)
301
+ (set! xxml_word_features (sable_pop_word_features))
302
+ UTT)
303
+ ("(PITCH" (ATTLIST UTT)
304
+ ;; Status: probably complete
305
+ ;; At present festival requires an utterance break here
306
+ (xxml_synth UTT)
307
+ (set! sable_pitch_context (cons int_lr_params sable_pitch_context))
308
+ (let ((base (sable_interpret_param
309
+ (car (xxml_attval "BASE" ATTLIST))
310
+ sable_pitch_base_map
311
+ (cadr (assoc 'target_f0_mean int_lr_params))
312
+ sable_pitch_base_original))
313
+ (med (sable_interpret_param
314
+ (car (xxml_attval "MED" ATTLIST))
315
+ sable_pitch_med_map
316
+ (cadr (assoc 'target_f0_mean int_lr_params))
317
+ sable_pitch_med_original))
318
+ (range (sable_interpret_param
319
+ (car (xxml_attval "RANGE" ATTLIST))
320
+ sable_pitch_range_map
321
+ (cadr (assoc 'target_f0_std int_lr_params))
322
+ sable_pitch_range_original))
323
+ (oldmean (cadr (assoc 'target_f0_mean int_lr_params))))
324
+ ;; Festival (if it supports anything) supports mean and std
325
+ ;; so we treat base as med if med doesn't seem to do anything
326
+ (if (equal? med oldmean)
327
+ (set! med base))
328
+ (set! int_lr_params
329
+ (cons
330
+ (list 'target_f0_mean med)
331
+ (cons
332
+ (list 'target_f0_std range)
333
+ int_lr_params)))
334
+ nil))
335
+ (")PITCH" (ATTLIST UTT)
336
+ (xxml_synth UTT)
337
+ (set! int_lr_params (car sable_pitch_context))
338
+ (set! sable_pitch_context (cdr sable_pitch_context))
339
+ nil)
340
+ ("(RATE" (ATTLIST UTT)
341
+ ;; Status: can't deal with absolute word per minute SPEED.
342
+ (sable_push_word_features)
343
+ ;; can't deal with words per minute value
344
+ (let ((rate (sable_interpret_param
345
+ (car (xxml_attval "SPEED" ATTLIST))
346
+ sable_rate_speed_map
347
+ (sable_find_fval "dur_stretch" xxml_word_features 1.0)
348
+ sable_rate_speed_original)))
349
+ (set! xxml_word_features
350
+ (cons (list "dur_stretch" (/ 1.0 rate)) xxml_word_features))
351
+ UTT))
352
+ (")RATE" (ATTLIST UTT)
353
+ (set! xxml_word_features (sable_pop_word_features))
354
+ UTT)
355
+ ("(VOLUME" (ATTLIST UTT)
356
+ ;; Status: probably complete
357
+ ;; At present festival requires an utterance break here
358
+ (xxml_synth UTT)
359
+ (set! sable_vol_context (cons (list sable_vol_type sable_vol_factor)
360
+ sable_vol_context))
361
+ (let ((level (sable_interpret_param
362
+ (car (xxml_attval "LEVEL" ATTLIST))
363
+ sable_volume_level_map
364
+ sable_vol_factor
365
+ 1.0)))
366
+ (cond
367
+ ((string-matches (car (xxml_attval "LEVEL" ATTLIST)) ".*%")
368
+ (set! sable_vol_type 'relative))
369
+ ((string-matches (car (xxml_attval "LEVEL" ATTLIST)) SABLE_RXDOUBLE)
370
+ (set! sable_vol_type 'absolute))
371
+ (t
372
+ (set! sable_vol_type 'relative)))
373
+ (set! sable_vol_factor level))
374
+ nil)
375
+ (")VOLUME" (ATTLIST UTT)
376
+ (xxml_synth UTT)
377
+ (set! sable_vol_type (car (car sable_vol_context)))
378
+ (set! sable_vol_factor (car (cdr (car sable_vol_context))))
379
+ (set! sable_vol_context (cdr sable_vol_context))
380
+ nil)
381
+ ("(ENGINE" (ATTLIST UTT)
382
+ ;; Status: probably complete
383
+ (xxml_synth UTT)
384
+ (if (string-matches (car (xxml_attval "ID" ATTLIST)) "festival.*")
385
+ (let ((datastr ""))
386
+ (mapcar
387
+ (lambda (c) (set! datastr (string-append datastr " " c)))
388
+ (xxml_attval "DATA" ATTLIST))
389
+ (apply_hooks tts_hooks (eval (list 'Utterance 'Text datastr)))
390
+ (set! sable_omitted_mode t)) ;; ignore contents
391
+ ;; else
392
+ ;; its not relevant to me
393
+ )
394
+ nil)
395
+ (")ENGINE" (ATTLIST UTT)
396
+ (xxml_synth UTT)
397
+ (set! sable_omitted_mode nil)
398
+ nil)
399
+ ("MARKER" (ATTLIST UTT)
400
+ ;; Status: does nothing
401
+ ;; Can't support this without low-level control of audio spooler
402
+ (format t "SABLE: marker \"%s\"\n"
403
+ (car (xxml_attval "MARK" ATTLIST)))
404
+ UTT)
405
+ ("(PRON" (ATTLIST UTT)
406
+ ;; Status: IPA currently ignored
407
+ (sable_push_word_features)
408
+ ;; can't deal with words per minute value
409
+ (let ((ipa (xxml_attval "IPA" ATTLIST))
410
+ (sub (xxml_attval "SUB" ATTLIST)))
411
+ (cond
412
+ (ipa
413
+ (format t "SABLE: ipa ignored\n")
414
+ (set! xxml_word_features
415
+ (cons (list "sable_ignore" "1") xxml_word_features)))
416
+ (sub
417
+ (set! xxml_word_features
418
+ (cons (list "sable_sub" (format nil "%l" sub))
419
+ xxml_word_features))
420
+ (set! xxml_word_features
421
+ (cons (list "sable_ignore" "1") xxml_word_features))))
422
+ UTT))
423
+ (")PRON" (ATTLIST UTT)
424
+ (set! xxml_word_features (sable_pop_word_features))
425
+ UTT)
426
+ ("(SAYAS" (ATTLIST UTT)
427
+ ;; Status: only a few of the types are dealt with
428
+ (sable_push_word_features)
429
+ (set! sable_utt UTT)
430
+ ;; can't deal with words per minute value
431
+ (let ((mode (downcase (car (xxml_attval "MODE" ATTLIST))))
432
+ (modetype (car (xxml_attval "MODETYPE" ATTLIST))))
433
+ (cond
434
+ ((string-equal mode "literal")
435
+ (set! xxml_word_features
436
+ (cons (list "sable_literal" "1") xxml_word_features)))
437
+ ((string-equal mode "phone")
438
+ (set! xxml_word_features
439
+ (cons (list "token_pos" "digits") xxml_word_features)))
440
+ ((string-equal mode "ordinal")
441
+ (set! xxml_word_features
442
+ (cons (list "token_pos" "ordinal") xxml_word_features)))
443
+ ((string-equal mode "cardinal")
444
+ (set! xxml_word_features
445
+ (cons (list "token_pos" "cardinal") xxml_word_features)))
446
+ (t
447
+ ;; blindly trust festival to get it right
448
+ t))
449
+ UTT))
450
+ (")SAYAS" (ATTLIST UTT)
451
+ (set! xxml_word_features (sable_pop_word_features))
452
+ UTT)
453
+
454
+
455
+ ))
456
+
457
+ (define (sable_init_func)
458
+ "(sable_init_func)
459
+ Initialisation for SABLE mode"
460
+ (sable_init_globals)
461
+ (voice_kal_diphone)
462
+ (set! sable_previous_elements xxml_elements)
463
+ (set! xxml_elements sable_elements)
464
+ (set! sable_previous_token_to_words english_token_to_words)
465
+ (set! english_token_to_words sable_token_to_words)
466
+ (set! token_to_words sable_token_to_words))
467
+
468
+ (define (sable_exit_func)
469
+ "(sable_exit_func)
470
+ Exit function for SABLE mode"
471
+ (set! xxml_elements sable_previous_elements)
472
+ (set! token_to_words sable_previous_token_to_words)
473
+ (set! english_token_to_words sable_previous_token_to_words))
474
+
475
+ (define (sable_push_word_features)
476
+ "(sable_push_word_features)
477
+ Save current word features on stack."
478
+ (set! sable_word_features_stack
479
+ (cons xxml_word_features sable_word_features_stack)))
480
+
481
+ (define (sable_adjust_volume utt)
482
+ "(sable_adjust_volume utt)
483
+ Amplify or attenutate signale based on value of sable_vol_factor
484
+ and sable_vol_type (absolute or relative)."
485
+ (set! utts (cons utt utts))
486
+ (cond
487
+ ((equal? sable_vol_type 'no_change)
488
+ utt)
489
+ ((equal? sable_vol_type 'absolute)
490
+ (utt.wave.rescale utt sable_vol_factor 'absolute))
491
+ ((equal? sable_vol_type 'relative)
492
+ (utt.wave.rescale utt sable_vol_factor))
493
+ (t
494
+ (format stderr "SABLE: volume unknown type \"%s\"\n" sable_vol_type)
495
+ utt))
496
+ utt)
497
+
498
+ (define (sable_pop_word_features)
499
+ "(sable_pop_word_features)
500
+ Pop word features from stack."
501
+ (let ((r (car sable_word_features_stack)))
502
+ (set! sable_word_features_stack (cdr sable_word_features_stack))
503
+ r))
504
+
505
+ (define (sable_find_fval feat flist def)
506
+ (cond
507
+ ((null flist) def)
508
+ ((string-equal feat (car (car flist)))
509
+ (car (cdr (car flist))))
510
+ (t
511
+ (sable_find_fval feat (cdr flist) def))))
512
+
513
+ (define (sable_interpret_param ident map original current)
514
+ "(sable_interpret_param IDENT MAP ORIGINAL CURRENT)
515
+ If IDENT is in map return ORIGINAL times value in map, otherwise
516
+ treat IDENT of the form +/-N% and modify CURRENT accordingly."
517
+ (let ((mm (assoc ident map)))
518
+ (cond
519
+ (mm
520
+ (* original (car (cdr mm))))
521
+ ((string-matches ident SABLE_RXDOUBLE)
522
+ (parse-number ident))
523
+ ((string-matches ident ".*%")
524
+ (+ current (* current (/ (parse-number (string-before ident "%"))
525
+ 100.0))))
526
+ ;; ((string-matches ident ".*%")
527
+ ;; (* current (/ (parse-number (string-before ident "%")) 100.0)))
528
+ ((not ident) current)
529
+ (t
530
+ (format stderr "SABLE: modifier \"%s\" not of float, tag or +/-N\n"
531
+ ident)
532
+ current))))
533
+
534
+ (define (sable_setup_voice_params)
535
+ "(sable_setup_voice_params)
536
+ Set up original values for various voice parameters."
537
+ (set! sable_pitch_base_original (cadr (assoc 'target_f0_mean int_lr_params)))
538
+ (set! sable_pitch_med_original (cadr (assoc 'target_f0_mean int_lr_params)))
539
+ (set! sable_pitch_range_original (cadr (assoc 'target_f0_std int_lr_params)))
540
+ (set! sable_rate_speed_original 1.0)
541
+ (if (and after_synth_hooks (not (consp after_synth_hooks)))
542
+ (set! after_synth_hooks
543
+ (cons after_synth_hooks (list sable_adjust_volume)))
544
+ (set! after_synth_hooks
545
+ (append after_synth_hooks (list sable_adjust_volume))))
546
+ )
547
+
548
+ ;;; Declare the new mode to Festival
549
+ (set! tts_text_modes
550
+ (cons
551
+ (list
552
+ 'sable ;; mode name
553
+ (list
554
+ (list 'init_func sable_init_func)
555
+ (list 'exit_func sable_exit_func)
556
+ '(analysis_type xml)
557
+ ))
558
+ tts_text_modes))
559
+
560
+ (provide 'sable-mode)
CosyVoice-ttsfrd/resource/festival/scfg.scm ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Some functions for manipulating a SCFG parse tree
35
+
36
+ (require_module 'parser)
37
+
38
+ (define (scfg_simplify tree)
39
+ "(scfg_brackets_only tree)
40
+ Output only the bracketing and the bottom level pos and words."
41
+ (cond
42
+ ((not tree) nil)
43
+ ((car (cdr (assoc 'pos (car (cdr (car tree))))))
44
+ ;; terminal node
45
+ (list
46
+ (car (cdr (assoc 'pos (car (cdr (car tree))))))
47
+ (car (car tree))))
48
+ (t
49
+ (cons
50
+ (car (car tree))
51
+ (mapcar scfg_simplify (cdr tree))))))
52
+
53
+ (define (scfg_simplify_relation_tree trees)
54
+ (mapcar scfg_simplify trees))
55
+
56
+ (defvar scfg_eos_tree eou_tree
57
+ "scfg_eos_tree
58
+ In MultiProbParse this CART tree is used to define end of sentence
59
+ within an utterance. It is applied to the token relation.
60
+ By default it is set to eou_tree.")
61
+
62
+ (provide 'scfg)
CosyVoice-ttsfrd/resource/festival/scfg_wsj_wp20.gram ADDED
@@ -0,0 +1,523 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;-*-mode:scheme-*-
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; A Stochastic context free grammar for the wp20 tag set with 19
35
+ ;;; nonterminals
36
+ ;;;
37
+ ;;; This was trained from 10,000 sentences (00-04) of the UPenn WSJ tree
38
+ ;;; bank using the inside-outside algorithm seeded with the bracketing from
39
+ ;;; the treebank. The implementation is the scfg_ suite in the
40
+ ;; speech tools and is based on the paper "Inside-Outside
41
+ ;;; Reestimation from partially bracketed corpora", F Pereira and
42
+ ;;; Y. Schabes. pp 128-135, 30th ACL, Newark, Delaware 1992.
43
+ ;;;
44
+ ;;; This grammar with 19 nonterminals was trained for 174 passes
45
+ ;;; using a fifth of training data each time. It was tested against
46
+ ;;; independent data both bracketed and unbracketed. After training,
47
+ ;;; all rules with a probability less than 1.0e-6 were pruned.
48
+ ;;;
49
+ ;;; On an unseen test set of 686 sentences (from wsj/05/) this gets
50
+ ;;; 92.2397% bracketing accuracy and 29.5918% sentences fully correct
51
+ ;;;
52
+ ;;; previous best 15_20 grammar
53
+ ;;; 90.2377% bracketing accuracy and 24.7813% sentences fully correct
54
+ ;;;
55
+ ;;; Training this grammar took a long time. This is best grammar
56
+ ;;; by testing grammars varying the number of non-terminals from 11-25
57
+ ;;; as the number of NTs increases the time for training also increases
58
+ ;;; This 19_20 grammar took 20 days on a Sun Ultra 1 140, but I also
59
+ ;;; had to search 11-18 to confirm this is best, which was done with a
60
+ ;;; collection of Ultra 140s 170s and Pentium Pros (Linux and FreeBSD)
61
+ ;;;
62
+ (0.00593452 NT00 NT00 NT00)
63
+ (0.0319023 NT00 NT00 NT13)
64
+ (0.00105452 NT00 NT00 NT18)
65
+ (0.00061816 NT00 NT02 NT10)
66
+ (0.000399698 NT00 NT02 NT12)
67
+ (0.0383818 NT00 NT05 NT00)
68
+ (0.00011458 NT00 NT06 NT03)
69
+ (0.00164298 NT00 NT06 NT17)
70
+ (0.00153884 NT00 NT07 NT07)
71
+ (0.00118244 NT00 NT07 NT12)
72
+ (0.00171642 NT00 NT07 NT13)
73
+ (0.00031308 NT00 NT07 NT17)
74
+ (0.0949408 NT00 NT09 NT18)
75
+ (0.000932166 NT00 NT10 NT03)
76
+ (0.000150288 NT00 NT10 NT17)
77
+ (0.0152371 NT00 NT12 NT18)
78
+ (0.73409 NT00 NT14 NT13)
79
+ (0.0403652 NT00 NT14 NT18)
80
+ (0.000195643 NT00 NT16 NT07)
81
+ (0.0134222 NT00 NT18 NT13)
82
+ (0.015624 NT00 NT18 NT18)
83
+ (0.00251118 NT01 NT01 NT07)
84
+ (0.00354571 NT01 NT01 NT11)
85
+ (0.22337 NT01 NT01 NT16)
86
+ (0.0467048 NT01 NT02 NT05)
87
+ (0.000518329 NT01 NT04 NT01)
88
+ (0.000100574 NT01 NT06 NT05)
89
+ (0.0480904 NT01 NT07 NT05)
90
+ (0.000358197 NT01 NT11 NT11)
91
+ (0.00278007 NT01 NT16 NT05)
92
+ (0.000179198 NT01 NT16 NT15)
93
+ (0.00140099 NT01 n)
94
+ (0.00228587 NT01 v)
95
+ (0.524988 NT01 dt)
96
+ (0.00128028 NT01 in)
97
+ (0.0660845 NT01 j)
98
+ (0.0131026 NT01 cd)
99
+ (0.00584238 NT01 r)
100
+ (0.0548382 NT01 prp)
101
+ (0.000445004 NT01 wdt)
102
+ (0.00135794 NT01 wp)
103
+ (0.000195991 NT01 wrb)
104
+ (0.000264526 NT02 NT01 NT01)
105
+ (0.00243627 NT02 NT01 NT02)
106
+ (0.613543 NT02 NT01 NT07)
107
+ (0.00180865 NT02 NT01 NT11)
108
+ (0.0042804 NT02 NT01 NT16)
109
+ (0.0392418 NT02 NT02 NT07)
110
+ (0.026104 NT02 NT02 NT12)
111
+ (0.000916683 NT02 NT02 NT16)
112
+ (0.00158862 NT02 NT04 NT01)
113
+ (0.000206161 NT02 NT04 NT02)
114
+ (0.00343189 NT02 NT04 NT16)
115
+ (0.000417113 NT02 NT07 NT05)
116
+ (0.0988457 NT02 NT07 NT07)
117
+ (0.000931386 NT02 NT07 NT11)
118
+ (0.00073236 NT02 NT07 NT12)
119
+ (0.000153421 NT02 NT10 NT13)
120
+ (0.00163484 NT02 NT11 NT02)
121
+ (0.0379562 NT02 NT11 NT07)
122
+ (0.0149 NT02 NT11 NT11)
123
+ (0.00105811 NT02 NT11 NT12)
124
+ (0.000175184 NT02 NT16 NT02)
125
+ (0.0403395 NT02 NT16 NT07)
126
+ (0.00297703 NT02 NT16 NT12)
127
+ (0.0875026 NT02 n)
128
+ (0.00496719 NT02 v)
129
+ (0.000409658 NT02 dt)
130
+ (0.00239978 NT02 j)
131
+ (0.010203 NT02 r)
132
+ (0.000194628 NT02 pdt)
133
+ (0.000377009 NT03 NT04 NT02)
134
+ (0.11551 NT03 NT08 NT13)
135
+ (0.347629 NT03 NT09 NT13)
136
+ (0.484911 NT03 NT10 NT13)
137
+ (0.00188291 NT03 NT11 NT12)
138
+ (0.0495461 NT03 NT17 NT13)
139
+ (0.00918797 NT04 NT03 NT05)
140
+ (0.000303954 NT04 NT04 NT02)
141
+ (0.00284848 NT04 NT04 NT04)
142
+ (0.00710115 NT04 NT04 NT12)
143
+ (0.000597744 NT04 NT04 NT15)
144
+ (0.000377075 NT04 NT04 NT16)
145
+ (0.00130088 NT04 NT09 NT05)
146
+ (0.00175428 NT04 NT10 NT13)
147
+ (0.000127716 NT04 NT15 NT04)
148
+ (0.00013648 NT04 NT15 NT06)
149
+ (0.00045093 NT04 NT15 NT07)
150
+ (0.000626479 NT04 NT15 NT16)
151
+ (0.000563588 NT04 NT16 NT15)
152
+ (0.0232089 NT04 NT17 NT05)
153
+ (0.000138094 NT04 NT17 NT15)
154
+ (0.00094009 NT04 n)
155
+ (0.671108 NT04 v)
156
+ (0.0150619 NT04 punc)
157
+ (0.00056566 NT04 dt)
158
+ (0.144629 NT04 r)
159
+ (0.00270621 NT04 prp)
160
+ (0.0449587 NT04 to)
161
+ (0.0543755 NT04 md)
162
+ (0.00839747 NT04 wdt)
163
+ (0.00813689 NT04 wp)
164
+ (0.000560496 NT05 NT07 NT05)
165
+ (0.000901219 NT05 NT15 NT07)
166
+ (0.180172 NT05 punc)
167
+ (0.533041 NT05 cc)
168
+ (0.285244 NT05 pos)
169
+ (0.00164003 NT06 NT00 NT13)
170
+ (0.00222915 NT06 NT01 NT06)
171
+ (0.275903 NT06 NT01 NT07)
172
+ (0.00191616 NT06 NT01 NT11)
173
+ (0.00316549 NT06 NT01 NT12)
174
+ (0.000730143 NT06 NT01 NT14)
175
+ (0.000559842 NT06 NT02 NT06)
176
+ (0.0236744 NT06 NT02 NT07)
177
+ (0.00284929 NT06 NT02 NT09)
178
+ (0.155052 NT06 NT02 NT12)
179
+ (0.00387995 NT06 NT02 NT14)
180
+ (0.0161403 NT06 NT02 NT18)
181
+ (0.000110944 NT06 NT04 NT01)
182
+ (0.00237845 NT06 NT04 NT02)
183
+ (0.00625142 NT06 NT04 NT06)
184
+ (0.00118802 NT06 NT04 NT08)
185
+ (0.000132901 NT06 NT04 NT10)
186
+ (0.000192545 NT06 NT04 NT11)
187
+ (0.000199118 NT06 NT06 NT01)
188
+ (0.0081704 NT06 NT06 NT12)
189
+ (0.00198439 NT06 NT06 NT14)
190
+ (0.000889455 NT06 NT06 NT18)
191
+ (0.00142038 NT06 NT07 NT05)
192
+ (0.0820095 NT06 NT07 NT07)
193
+ (0.000112894 NT06 NT07 NT09)
194
+ (0.0220243 NT06 NT07 NT12)
195
+ (0.000133911 NT06 NT07 NT14)
196
+ (0.00100807 NT06 NT07 NT17)
197
+ (0.000191764 NT06 NT08 NT13)
198
+ (0.000340112 NT06 NT10 NT08)
199
+ (0.000126776 NT06 NT10 NT09)
200
+ (0.0136266 NT06 NT10 NT12)
201
+ (0.00867414 NT06 NT10 NT13)
202
+ (0.00341334 NT06 NT10 NT18)
203
+ (0.00154851 NT06 NT11 NT12)
204
+ (0.00104947 NT06 NT12 NT12)
205
+ (0.000219189 NT06 NT14 NT05)
206
+ (0.00313879 NT06 NT14 NT13)
207
+ (0.000745073 NT06 NT15 NT02)
208
+ (0.000433144 NT06 NT15 NT06)
209
+ (0.000159867 NT06 NT15 NT16)
210
+ (0.00124313 NT06 NT16 NT02)
211
+ (0.00918606 NT06 NT16 NT07)
212
+ (0.00373496 NT06 NT16 NT12)
213
+ (0.014053 NT06 NT18 NT13)
214
+ (0.0155714 NT06 n)
215
+ (0.00123379 NT06 punc)
216
+ (0.0152764 NT06 dt)
217
+ (0.00123486 NT06 j)
218
+ (0.00359625 NT06 r)
219
+ (0.212966 NT06 prp)
220
+ (0.00199168 NT06 cc)
221
+ (0.0383471 NT06 wdt)
222
+ (0.0182587 NT06 wp)
223
+ (0.00204833 NT06 wrb)
224
+ (0.0109929 NT06 ex)
225
+ (0.0011995 NT07 NT05 NT16)
226
+ (0.119588 NT07 NT07 NT07)
227
+ (0.000353596 NT07 NT07 NT11)
228
+ (0.000177793 NT07 NT07 NT12)
229
+ (0.00101956 NT07 NT11 NT11)
230
+ (0.000357614 NT07 NT15 NT01)
231
+ (0.00084812 NT07 NT15 NT06)
232
+ (0.0182872 NT07 NT16 NT07)
233
+ (0.00018607 NT07 NT16 NT11)
234
+ (0.856315 NT07 n)
235
+ (0.000736333 NT07 v)
236
+ (0.000645479 NT08 NT00 NT09)
237
+ (0.000990156 NT08 NT01 NT02)
238
+ (0.0410251 NT08 NT01 NT07)
239
+ (0.0013863 NT08 NT01 NT09)
240
+ (0.000242552 NT08 NT01 NT12)
241
+ (0.00174478 NT08 NT01 NT14)
242
+ (0.000596656 NT08 NT01 NT16)
243
+ (0.00130945 NT08 NT02 NT07)
244
+ (0.166303 NT08 NT02 NT09)
245
+ (0.0143253 NT08 NT02 NT12)
246
+ (0.0113813 NT08 NT02 NT14)
247
+ (0.000597887 NT08 NT02 NT16)
248
+ (0.0133053 NT08 NT03 NT09)
249
+ (0.0109076 NT08 NT03 NT17)
250
+ (0.000211313 NT08 NT04 NT01)
251
+ (0.0105796 NT08 NT04 NT02)
252
+ (0.00440181 NT08 NT04 NT04)
253
+ (0.00203737 NT08 NT04 NT06)
254
+ (0.213275 NT08 NT04 NT08)
255
+ (0.0781169 NT08 NT04 NT09)
256
+ (0.0190657 NT08 NT04 NT10)
257
+ (0.00319326 NT08 NT04 NT12)
258
+ (0.000693766 NT08 NT04 NT15)
259
+ (0.00112226 NT08 NT04 NT16)
260
+ (0.00117025 NT08 NT06 NT02)
261
+ (0.00807496 NT08 NT06 NT08)
262
+ (0.0183971 NT08 NT06 NT09)
263
+ (0.00127343 NT08 NT06 NT14)
264
+ (0.0322725 NT08 NT06 NT17)
265
+ (0.00396897 NT08 NT07 NT07)
266
+ (0.0154729 NT08 NT07 NT09)
267
+ (0.000708139 NT08 NT07 NT10)
268
+ (0.00186499 NT08 NT07 NT11)
269
+ (0.000701346 NT08 NT07 NT14)
270
+ (0.0116278 NT08 NT08 NT09)
271
+ (0.0965117 NT08 NT10 NT09)
272
+ (0.000142086 NT08 NT10 NT12)
273
+ (0.000210725 NT08 NT10 NT14)
274
+ (0.00336223 NT08 NT11 NT07)
275
+ (0.00183799 NT08 NT11 NT09)
276
+ (0.00109249 NT08 NT11 NT11)
277
+ (0.000880671 NT08 NT11 NT12)
278
+ (0.0032493 NT08 NT12 NT08)
279
+ (0.0372072 NT08 NT12 NT09)
280
+ (0.00113127 NT08 NT12 NT12)
281
+ (0.00892231 NT08 NT15 NT02)
282
+ (0.00383754 NT08 NT15 NT06)
283
+ (0.000528365 NT08 NT15 NT07)
284
+ (0.0060705 NT08 NT15 NT08)
285
+ (0.00853698 NT08 NT15 NT10)
286
+ (0.0349777 NT08 NT15 NT14)
287
+ (0.000202857 NT08 NT16 NT06)
288
+ (0.00709689 NT08 NT16 NT07)
289
+ (0.000240097 NT08 NT16 NT08)
290
+ (0.0401819 NT08 NT16 NT09)
291
+ (0.00124754 NT08 NT16 NT14)
292
+ (0.00862498 NT08 n)
293
+ (0.0115193 NT08 v)
294
+ (0.000974267 NT08 in)
295
+ (0.0169837 NT08 j)
296
+ (0.00626434 NT08 r)
297
+ (0.00437851 NT08 prp)
298
+ (0.0062359 NT09 NT01 NT07)
299
+ (0.000165196 NT09 NT01 NT14)
300
+ (0.00151872 NT09 NT02 NT04)
301
+ (0.000660061 NT09 NT02 NT15)
302
+ (0.000434321 NT09 NT02 NT16)
303
+ (0.00805872 NT09 NT03 NT09)
304
+ (0.000180982 NT09 NT04 NT08)
305
+ (0.050609 NT09 NT04 NT09)
306
+ (0.000307442 NT09 NT04 NT15)
307
+ (0.00281491 NT09 NT04 NT17)
308
+ (0.000295911 NT09 NT06 NT15)
309
+ (0.00133828 NT09 NT07 NT11)
310
+ (0.0235741 NT09 NT12 NT09)
311
+ (0.00121997 NT09 NT12 NT12)
312
+ (0.00391762 NT09 NT15 NT01)
313
+ (0.173027 NT09 NT15 NT02)
314
+ (0.000462089 NT09 NT15 NT06)
315
+ (0.0276663 NT09 NT15 NT07)
316
+ (0.210483 NT09 NT15 NT08)
317
+ (0.000177004 NT09 NT15 NT09)
318
+ (0.243402 NT09 NT15 NT10)
319
+ (0.0174403 NT09 NT15 NT11)
320
+ (0.00646962 NT09 NT15 NT12)
321
+ (0.155174 NT09 NT15 NT14)
322
+ (0.00930502 NT09 NT15 NT17)
323
+ (0.000311399 NT09 NT16 NT02)
324
+ (0.0052031 NT09 NT16 NT07)
325
+ (0.00742336 NT09 NT16 NT09)
326
+ (0.000409254 NT09 in)
327
+ (0.0019424 NT09 j)
328
+ (0.0393282 NT09 r)
329
+ (0.00016039 NT09 prp)
330
+ (0.00268682 NT10 NT01 NT07)
331
+ (0.00173594 NT10 NT01 NT09)
332
+ (0.00550051 NT10 NT01 NT10)
333
+ (0.00269002 NT10 NT01 NT11)
334
+ (0.00881491 NT10 NT01 NT12)
335
+ (0.0158503 NT10 NT02 NT02)
336
+ (0.00229071 NT10 NT02 NT07)
337
+ (0.00765082 NT10 NT02 NT09)
338
+ (0.00102327 NT10 NT02 NT11)
339
+ (0.474288 NT10 NT02 NT12)
340
+ (0.0119086 NT10 NT02 NT14)
341
+ (0.000270767 NT10 NT02 NT15)
342
+ (0.00425023 NT10 NT02 NT16)
343
+ (0.0533347 NT10 NT04 NT02)
344
+ (0.00286524 NT10 NT04 NT06)
345
+ (0.0687658 NT10 NT04 NT10)
346
+ (0.0157381 NT10 NT04 NT12)
347
+ (0.000809508 NT10 NT05 NT12)
348
+ (0.00188343 NT10 NT06 NT04)
349
+ (0.000155481 NT10 NT06 NT09)
350
+ (0.00569591 NT10 NT06 NT14)
351
+ (0.00233367 NT10 NT06 NT17)
352
+ (0.000189475 NT10 NT07 NT05)
353
+ (0.018548 NT10 NT07 NT07)
354
+ (0.00472354 NT10 NT07 NT09)
355
+ (0.0121145 NT10 NT07 NT11)
356
+ (0.0698482 NT10 NT07 NT12)
357
+ (0.000402661 NT10 NT07 NT16)
358
+ (0.00183044 NT10 NT07 NT17)
359
+ (0.00166519 NT10 NT10 NT02)
360
+ (0.015445 NT10 NT10 NT09)
361
+ (0.019208 NT10 NT10 NT12)
362
+ (0.000942866 NT10 NT10 NT18)
363
+ (0.00149941 NT10 NT11 NT01)
364
+ (0.00624706 NT10 NT11 NT02)
365
+ (0.0381755 NT10 NT11 NT11)
366
+ (0.00754256 NT10 NT11 NT12)
367
+ (0.00139213 NT10 NT15 NT02)
368
+ (0.000523505 NT10 NT15 NT06)
369
+ (0.0015256 NT10 NT15 NT10)
370
+ (0.00119525 NT10 NT15 NT12)
371
+ (0.00683524 NT10 NT16 NT02)
372
+ (0.000398591 NT10 NT16 NT04)
373
+ (0.0701558 NT10 NT16 NT07)
374
+ (0.00198721 NT10 NT16 NT11)
375
+ (0.0075364 NT10 NT16 NT12)
376
+ (0.0186618 NT10 n)
377
+ (0.000591828 NT10 uh)
378
+ (0.157827 NT11 NT11 NT11)
379
+ (0.0422576 NT11 NT15 NT11)
380
+ (0.00247895 NT11 NT15 NT16)
381
+ (0.000257833 NT11 dt)
382
+ (0.754818 NT11 cd)
383
+ (0.0421123 NT11 r)
384
+ (0.00236916 NT12 NT01 NT07)
385
+ (0.000118511 NT12 NT02 NT16)
386
+ (0.00638739 NT12 NT04 NT02)
387
+ (0.0055731 NT12 NT04 NT04)
388
+ (0.0340903 NT12 NT04 NT12)
389
+ (0.00102031 NT12 NT04 NT15)
390
+ (0.00143793 NT12 NT04 NT16)
391
+ (0.000102621 NT12 NT04 NT17)
392
+ (0.0032774 NT12 NT06 NT04)
393
+ (0.000366976 NT12 NT07 NT07)
394
+ (0.00218153 NT12 NT07 NT11)
395
+ (0.0117989 NT12 NT11 NT07)
396
+ (0.00303601 NT12 NT12 NT12)
397
+ (0.0747798 NT12 NT13 NT03)
398
+ (0.000232806 NT12 NT15 NT01)
399
+ (0.341016 NT12 NT15 NT02)
400
+ (0.0190932 NT12 NT15 NT06)
401
+ (0.100931 NT12 NT15 NT07)
402
+ (0.193386 NT12 NT15 NT10)
403
+ (0.0142796 NT12 NT15 NT11)
404
+ (0.000915196 NT12 NT16 NT07)
405
+ (0.000299768 NT12 NT16 NT11)
406
+ (0.0135637 NT12 NT16 NT12)
407
+ (0.115493 NT12 n)
408
+ (0.00344871 NT12 v)
409
+ (0.0262404 NT12 punc)
410
+ (0.000493049 NT12 in)
411
+ (0.00235382 NT12 j)
412
+ (0.0192274 NT12 r)
413
+ (0.00199831 NT12 prp)
414
+ (0.000209376 NT13 NT11 NT15)
415
+ (0.00188858 NT13 NT13 NT03)
416
+ (0.540855 NT13 punc)
417
+ (0.00804226 NT13 cc)
418
+ (0.000413617 NT14 NT00 NT09)
419
+ (0.0218326 NT14 NT00 NT14)
420
+ (0.000451496 NT14 NT00 NT18)
421
+ (0.00149459 NT14 NT01 NT07)
422
+ (0.00384046 NT14 NT01 NT17)
423
+ (0.00138254 NT14 NT02 NT09)
424
+ (0.0525259 NT14 NT03 NT14)
425
+ (0.000893974 NT14 NT04 NT02)
426
+ (0.000175088 NT14 NT04 NT06)
427
+ (0.000478859 NT14 NT04 NT08)
428
+ (0.00086439 NT14 NT04 NT09)
429
+ (0.00529624 NT14 NT04 NT10)
430
+ (0.000476852 NT14 NT04 NT12)
431
+ (0.00549502 NT14 NT04 NT14)
432
+ (0.0281873 NT14 NT05 NT14)
433
+ (0.76715 NT14 NT06 NT17)
434
+ (0.00303311 NT14 NT07 NT07)
435
+ (0.00027137 NT14 NT07 NT09)
436
+ (0.000748841 NT14 NT07 NT12)
437
+ (0.0874896 NT14 NT07 NT17)
438
+ (0.00416962 NT14 NT09 NT14)
439
+ (0.00175999 NT14 NT10 NT09)
440
+ (0.000710869 NT14 NT11 NT17)
441
+ (0.000723932 NT14 NT12 NT07)
442
+ (0.00440147 NT14 NT12 NT14)
443
+ (0.000761726 NT14 NT14 NT09)
444
+ (0.00084762 NT14 NT14 NT17)
445
+ (0.000323644 NT14 NT15 NT02)
446
+ (0.00264492 NT14 NT15 NT14)
447
+ (0.000238841 NT14 NT16 NT07)
448
+ (0.000126025 NT14 NT16 NT09)
449
+ (0.000217731 NT14 r)
450
+ (0.00024161 NT14 wrb)
451
+ (0.000366989 NT15 NT04 NT04)
452
+ (0.00127143 NT15 NT04 NT15)
453
+ (0.00137902 NT15 NT11 NT07)
454
+ (0.000109067 NT15 NT15 NT04)
455
+ (0.00380199 NT15 NT15 NT06)
456
+ (0.000193842 NT15 NT15 NT15)
457
+ (0.000253898 NT15 NT15 NT16)
458
+ (0.00556123 NT15 v)
459
+ (0.0798535 NT15 punc)
460
+ (0.557206 NT15 in)
461
+ (0.0519477 NT15 cc)
462
+ (0.170466 NT15 of)
463
+ (0.113587 NT15 to)
464
+ (0.0125211 NT15 wrb)
465
+ (0.00146961 NT15 pdt)
466
+ (0.000682686 NT16 NT01 NT16)
467
+ (0.000353409 NT16 NT02 NT02)
468
+ (0.0034721 NT16 NT02 NT05)
469
+ (0.00392739 NT16 NT04 NT04)
470
+ (0.0225952 NT16 NT04 NT16)
471
+ (0.00368407 NT16 NT05 NT16)
472
+ (0.000275916 NT16 NT06 NT05)
473
+ (0.0263102 NT16 NT07 NT05)
474
+ (0.00344251 NT16 NT07 NT12)
475
+ (0.00271063 NT16 NT07 NT16)
476
+ (0.000950873 NT16 NT10 NT13)
477
+ (0.0229124 NT16 NT11 NT07)
478
+ (0.0173136 NT16 NT11 NT11)
479
+ (0.0094147 NT16 NT11 NT16)
480
+ (0.00210054 NT16 NT13 NT03)
481
+ (0.000417271 NT16 NT15 NT01)
482
+ (0.0100377 NT16 NT15 NT11)
483
+ (0.000679194 NT16 NT16 NT05)
484
+ (0.00203961 NT16 NT16 NT11)
485
+ (0.00352444 NT16 NT16 NT12)
486
+ (0.0133536 NT16 NT16 NT16)
487
+ (0.0041124 NT16 n)
488
+ (0.0518387 NT16 v)
489
+ (0.0133556 NT16 punc)
490
+ (0.746857 NT16 j)
491
+ (0.0325454 NT16 cd)
492
+ (0.000994964 NT16 r)
493
+ (0.000325555 NT17 NT03 NT09)
494
+ (0.000431668 NT17 NT03 NT17)
495
+ (0.000283523 NT17 NT04 NT01)
496
+ (0.00308221 NT17 NT04 NT02)
497
+ (0.000106449 NT17 NT04 NT07)
498
+ (0.584517 NT17 NT04 NT08)
499
+ (0.0389749 NT17 NT04 NT09)
500
+ (0.00927257 NT17 NT04 NT10)
501
+ (0.000698039 NT17 NT04 NT11)
502
+ (0.0594712 NT17 NT04 NT14)
503
+ (0.000381951 NT17 NT04 NT16)
504
+ (0.248255 NT17 NT04 NT17)
505
+ (0.000264379 NT17 NT05 NT08)
506
+ (0.00194384 NT17 NT05 NT10)
507
+ (0.000308808 NT17 NT05 NT14)
508
+ (0.000271388 NT17 NT07 NT08)
509
+ (0.000131093 NT17 NT07 NT10)
510
+ (0.00011195 NT17 NT07 NT17)
511
+ (0.000462643 NT17 NT08 NT09)
512
+ (0.00153331 NT17 NT11 NT07)
513
+ (0.00214335 NT17 NT11 NT11)
514
+ (0.000307068 NT17 NT11 NT12)
515
+ (0.000550528 NT17 NT15 NT10)
516
+ (0.000125644 NT17 NT16 NT02)
517
+ (0.000474489 NT17 NT17 NT09)
518
+ (0.00032483 NT17 NT17 NT18)
519
+ (0.045027 NT17 v)
520
+ (0.00425503 NT18 NT07 NT18)
521
+ (0.978831 NT18 NT13 NT00)
522
+ (0.00130119 NT18 NT13 NT03)
523
+ (0.0155958 NT18 NT17 NT13)
CosyVoice-ttsfrd/resource/festival/tobi.scm ADDED
@@ -0,0 +1,338 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; A CART tree for predicting ToBI accents (learned from f2b)
35
+ ;;; punctuation and minimal pos
36
+ ;;;
37
+
38
+ ; NON !H* L+H L*+
39
+ ; NONE10265 583 66 40 0 0 10954 [10265/10954] 93.710
40
+ ; H* 650 1805 61 57 0 0 2573 [1805/2573] 70.152
41
+ ; !H* 317 241 125 42 0 0 725 [125/725] 17.241
42
+ ; L+H* 457 486 76 80 0 0 1099 [80/1099] 7.279
43
+ ; L* 45 113 14 4 0 0 176 [0/176] 0.000
44
+ ; L*+H 6 6 0 1 0 0 13 [0/13] 0.000
45
+ ; 11740 3234 342 224 0 0
46
+ ;total 15540 correct 12275.000 78.990%
47
+
48
+ (set! f2b_int_accent_cart_tree
49
+ '
50
+ ;; these first few lines are hand written to deal with emphasis (from ssml)
51
+ ((R:SylStructure.parent.R:Token.parent.EMPH is 1)
52
+ (((NONE 0.0) (H* 1) (!H* 0.0) (L+H* 0) (L* 0) (L*+H 0) H*))
53
+ ((n.R:SylStructure.parent.R:Token.parent.EMPH is 1)
54
+ (((NONE 1.0) (H* 0) (!H* 0.0) (L+H* 0) (L* 0) (L*+H 0) NONE))
55
+ ((p.R:SylStructure.parent.R:Token.parent.EMPH is 1)
56
+ (((NONE 1.0) (H* 0) (!H* 0.0) (L+H* 0) (L* 0) (L*+H 0) NONE))
57
+
58
+ ((ssyl_in is 10)
59
+ (((NONE 0.99726) (H* 0) (!H* 0.00273973) (L+H* 0) (L* 0) (L*+H 0) NONE))
60
+ ((R:SylStructure.parent.gpos is to)
61
+ (((NONE 0.995984) (H* 0.00401606) (!H* 0) (L+H* 0) (L* 0) (L*+H 0) NONE))
62
+ ((R:SylStructure.parent.gpos is cc)
63
+ (((NONE 0.987768) (H* 0.00611621) (!H* 0) (L+H* 0.00611621) (L* 0) (L*+H 0) NONE))
64
+ ((ssyl_out is 10)
65
+ (((NONE 0.927273) (H* 0.0545455) (!H* 0) (L+H* 0.0181818) (L* 0) (L*+H 0) NONE))
66
+ ((R:SylStructure.parent.gpos is in)
67
+ (((NONE 0.938322) (H* 0.0353618) (!H* 0.00493421) (L+H* 0.0197368) (L* 0.00164474) (L*+H 0) NONE))
68
+ ((R:SylStructure.parent.gpos is wp)
69
+ (((NONE 0.895238) (H* 0.0857143) (!H* 0) (L+H* 0.0190476) (L* 0) (L*+H 0) NONE))
70
+ ((R:SylStructure.parent.gpos is aux)
71
+ (((NONE 0.912281) (H* 0.0380117) (!H* 0.00584795) (L+H* 0.0350877) (L* 0.00584795) (L*+H 0.00292398) NONE))
72
+ ((R:SylStructure.parent.gpos is det)
73
+ (((NONE 0.898004) (H* 0.0643016) (!H* 0.00332594) (L+H* 0.0332594) (L* 0) (L*+H 0.00110865) NONE))
74
+ ((stress is 0)
75
+ (((NONE 0.978415) (H* 0.0144999) (!H* 0.00164772) (L+H* 0.00510793) (L* 0.000329544) (L*+H 0) NONE))
76
+ ((R:SylStructure.parent.R:Word.p.gpos is 0)
77
+ (((NONE 0.209877) (H* 0.716049) (!H* 0) (L+H* 0.0617284) (L* 0.0123457) (L*+H 0) H*))
78
+ ((R:SylStructure.parent.gpos is md)
79
+ (((NONE 0.693548) (H* 0.177419) (!H* 0.0322581) (L+H* 0.0967742) (L* 0) (L*+H 0) NONE))
80
+ ((p.syl_break is 3)
81
+ ((syl_break is 1)
82
+ (((NONE 0.4375) (H* 0.416667) (!H* 0) (L+H* 0.135417) (L* 0.0104167) (L*+H 0) NONE))
83
+ (((NONE 0.171171) (H* 0.666667) (!H* 0) (L+H* 0.144144) (L* 0.018018) (L*+H 0) H*)))
84
+ ((pp.syl_break is 4)
85
+ ((R:SylStructure.parent.R:Word.pp.gpos is in)
86
+ (((NONE 0.0980392) (H* 0.803922) (!H* 0) (L+H* 0.0784314) (L* 0.0196078) (L*+H 0) H*))
87
+ ((syl_out is 0)
88
+ (((NONE 0.0185185) (H* 0.796296) (!H* 0.037037) (L+H* 0.0925926) (L* 0.0555556) (L*+H 0) H*))
89
+ ((R:SylStructure.parent.R:Word.n.gpos is in)
90
+ (((NONE 0.132353) (H* 0.676471) (!H* 0) (L+H* 0.161765) (L* 0.0294118) (L*+H 0) H*))
91
+ ((syl_break is 0)
92
+ (((NONE 0.125) (H* 0.633929) (!H* 0.0133929) (L+H* 0.183036) (L* 0.0401786) (L*+H 0.00446429) H*))
93
+ ((n.stress is 0)
94
+ (((NONE 0.364865) (H* 0.567568) (!H* 0) (L+H* 0.0540541) (L* 0.0135135) (L*+H 0) H*))
95
+ ((p.syl_break is 0)
96
+ (((NONE 0.612903) (H* 0.290323) (!H* 0) (L+H* 0.0967742) (L* 0) (L*+H 0) NONE))
97
+ (((NONE 0.32) (H* 0.44) (!H* 0.02) (L+H* 0.22) (L* 0) (L*+H 0) H*))))))))
98
+ ((ssyl_in is 0)
99
+ (((NONE 0.167769) (H* 0.628926) (!H* 0.0214876) (L+H* 0.142975) (L* 0.0363636) (L*+H 0.00247934) H*))
100
+ ((ssyl_out is 4)
101
+ (((NONE 0.490385) (H* 0.240385) (!H* 0.0961538) (L+H* 0.163462) (L* 0.00961538) (L*+H 0) NONE))
102
+ ((pp.syl_break is 3)
103
+ ((R:SylStructure.parent.R:Word.p.gpos is content)
104
+ (((NONE 0.346154) (H* 0.346154) (!H* 0.0769231) (L+H* 0.192308) (L* 0.0384615) (L*+H 0) NONE))
105
+ (((NONE 0.160714) (H* 0.571429) (!H* 0.0178571) (L+H* 0.178571) (L* 0.0714286) (L*+H 0) H*)))
106
+ ((syl_in is 2)
107
+ ((n.stress is 0)
108
+ ((R:SylStructure.parent.R:Word.p.gpos is in)
109
+ (((NONE 0.218182) (H* 0.618182) (!H* 0.0363636) (L+H* 0.0909091) (L* 0.0181818) (L*+H 0.0181818) H*))
110
+ ((syl_out is 2)
111
+ (((NONE 0.0961538) (H* 0.634615) (!H* 0.0961538) (L+H* 0.134615) (L* 0.0384615) (L*+H 0) H*))
112
+ ((R:SylStructure.parent.R:Word.p.gpos is content)
113
+ ((syl_out is 4)
114
+ (((NONE 0.56) (H* 0.12) (!H* 0.08) (L+H* 0.24) (L* 0) (L*+H 0) NONE))
115
+ (((NONE 0.262821) (H* 0.378205) (!H* 0.121795) (L+H* 0.192308) (L* 0.0448718) (L*+H 0) H*)))
116
+ (((NONE 0.161905) (H* 0.590476) (!H* 0.0285714) (L+H* 0.171429) (L* 0.047619) (L*+H 0) H*)))))
117
+ ((n.syl_break is 0)
118
+ (((NONE 0.551724) (H* 0.293103) (!H* 0) (L+H* 0.155172) (L* 0) (L*+H 0) NONE))
119
+ (((NONE 0.408451) (H* 0.422535) (!H* 0.056338) (L+H* 0.112676) (L* 0) (L*+H 0) H*))))
120
+ ((R:SylStructure.parent.R:Word.n.gpos is 0)
121
+ ((syl_break is 0)
122
+ (((NONE 0.105263) (H* 0.315789) (!H* 0.157895) (L+H* 0.421053) (L* 0) (L*+H 0) L+H*))
123
+ (((NONE 0.641509) (H* 0.132075) (!H* 0.132075) (L+H* 0.0943396) (L* 0) (L*+H 0) NONE)))
124
+ ((syl_break is 1)
125
+ ((ssyl_in is 3)
126
+ (((NONE 0.638889) (H* 0.152778) (!H* 0.125) (L+H* 0.0833333) (L* 0) (L*+H 0) NONE))
127
+ ((p.syl_break is 0)
128
+ (((NONE 0.551402) (H* 0.186916) (!H* 0.158879) (L+H* 0.0841122) (L* 0.0186916) (L*+H 0) NONE))
129
+ ((n.stress is 0)
130
+ ((pp.syl_break is 0)
131
+ (((NONE 0.413043) (H* 0.184783) (!H* 0.152174) (L+H* 0.23913) (L* 0.0108696) (L*+H 0) NONE))
132
+ (((NONE 0.2125) (H* 0.3375) (!H* 0.1875) (L+H* 0.2125) (L* 0.05) (L*+H 0) H*)))
133
+ (((NONE 0.449153) (H* 0.245763) (!H* 0.101695) (L+H* 0.20339) (L* 0) (L*+H 0) NONE)))))
134
+ ((syl_out is 4)
135
+ ((nn.syl_break is 0)
136
+ ((pp.syl_break is 0)
137
+ (((NONE 0.45614) (H* 0.210526) (!H* 0.192982) (L+H* 0.140351) (L* 0) (L*+H 0) NONE))
138
+ (((NONE 0.288462) (H* 0.25) (!H* 0.0961538) (L+H* 0.346154) (L* 0) (L*+H 0.0192308) L+H*)))
139
+ (((NONE 0.163934) (H* 0.459016) (!H* 0.131148) (L+H* 0.245902) (L* 0) (L*+H 0) H*)))
140
+ ((syl_out is 5)
141
+ ((R:SylStructure.parent.R:Word.p.gpos is content)
142
+ (((NONE 0.372881) (H* 0.20339) (!H* 0.169492) (L+H* 0.220339) (L* 0.0338983) (L*+H 0) NONE))
143
+ (((NONE 0.0961538) (H* 0.673077) (!H* 0.115385) (L+H* 0.0961538) (L* 0.0192308) (L*+H 0) H*)))
144
+ ((R:SylStructure.parent.R:Word.pp.gpos is in)
145
+ ((syl_in is 4)
146
+ (((NONE 0.352113) (H* 0.422535) (!H* 0.15493) (L+H* 0.0704225) (L* 0) (L*+H 0) H*))
147
+ ((syl_in is 3)
148
+ (((NONE 0.290323) (H* 0.467742) (!H* 0.0806452) (L+H* 0.145161) (L* 0.016129) (L*+H 0) H*))
149
+ ((pp.syl_break is 0)
150
+ (((NONE 0.465517) (H* 0.293103) (!H* 0.172414) (L+H* 0.0689655) (L* 0) (L*+H 0) NONE))
151
+ ((R:SylStructure.parent.R:Word.p.gpos is content)
152
+ (((NONE 0.18) (H* 0.36) (!H* 0.28) (L+H* 0.14) (L* 0.04) (L*+H 0) H*))
153
+ (((NONE 0.0877193) (H* 0.22807) (!H* 0.368421) (L+H* 0.298246) (L* 0.0175439) (L*+H 0) !H*))))))
154
+ ((ssyl_out is 2)
155
+ ((p.syl_break is 0)
156
+ (((NONE 0.634921) (H* 0.174603) (!H* 0.0793651) (L+H* 0.111111) (L* 0) (L*+H 0) NONE))
157
+ ((pp.syl_break is 0)
158
+ (((NONE 0.388889) (H* 0.148148) (!H* 0.148148) (L+H* 0.259259) (L* 0.0185185) (L*+H 0.037037) NONE))
159
+ (((NONE 0.294118) (H* 0.137255) (!H* 0.215686) (L+H* 0.333333) (L* 0.0196078) (L*+H 0) L+H*))))
160
+ ((R:SylStructure.parent.R:Word.pp.gpos is to)
161
+ (((NONE 0.0877193) (H* 0.350877) (!H* 0.210526) (L+H* 0.315789) (L* 0.0350877) (L*+H 0) H*))
162
+ ((syl_break is 3)
163
+ ((pp.syl_break is 0)
164
+ (((NONE 0.478261) (H* 0.141304) (!H* 0.195652) (L+H* 0.184783) (L* 0) (L*+H 0) NONE))
165
+ (((NONE 0.217822) (H* 0.366337) (!H* 0.257426) (L+H* 0.128713) (L* 0.029703) (L*+H 0) H*)))
166
+ ((syl_in is 7)
167
+ ((n.stress is 0)
168
+ ((R:SylStructure.parent.R:Word.n.gpos is content)
169
+ (((NONE 0.117647) (H* 0.220588) (!H* 0.441176) (L+H* 0.176471) (L* 0.0441176) (L*+H 0) !H*))
170
+ (((NONE 0.415385) (H* 0.0461538) (!H* 0.2) (L+H* 0.246154) (L* 0.0923077) (L*+H 0) NONE)))
171
+ (((NONE 0.716981) (H* 0.113208) (!H* 0.0943396) (L+H* 0.0754717) (L* 0) (L*+H 0) NONE)))
172
+ ((R:SylStructure.parent.R:Word.n.gpos is cc)
173
+ (((NONE 0.292308) (H* 0.184615) (!H* 0.276923) (L+H* 0.246154) (L* 0) (L*+H 0) NONE))
174
+ ((nn.syl_break is 3)
175
+ (((NONE 0.2) (H* 0.333333) (!H* 0.283333) (L+H* 0.15) (L* 0.0333333) (L*+H 0) H*))
176
+ ((ssyl_in is 4)
177
+ (((NONE 0.383838) (H* 0.151515) (!H* 0.212121) (L+H* 0.20202) (L* 0.050505) (L*+H 0) NONE))
178
+ ((p.syl_break is 0)
179
+ ((n.syl_break is 1)
180
+ (((NONE 0.526316) (H* 0.210526) (!H* 0.0921053) (L+H* 0.171053) (L* 0) (L*+H 0) NONE))
181
+ ((ssyl_in is 3)
182
+ (((NONE 0.509804) (H* 0.0980392) (!H* 0.215686) (L+H* 0.156863) (L* 0.0196078) (L*+H 0) NONE))
183
+ ((pp.syl_break is 0)
184
+ (((NONE 0.506667) (H* 0.173333) (!H* 0.106667) (L+H* 0.2) (L* 0.0133333) (L*+H 0) NONE))
185
+ ((ssyl_in is 1)
186
+ (((NONE 0.1) (H* 0.4) (!H* 0.266667) (L+H* 0.188889) (L* 0.0444444) (L*+H 0) H*))
187
+ (((NONE 0.326316) (H* 0.210526) (!H* 0.221053) (L+H* 0.189474) (L* 0.0526316) (L*+H 0) NONE))))))
188
+ ((R:SylStructure.parent.R:Word.p.gpos is in)
189
+ (((NONE 0.0625) (H* 0.296875) (!H* 0.265625) (L+H* 0.328125) (L* 0.046875) (L*+H 0) L+H*))
190
+ ((syl_in is 6)
191
+ (((NONE 0.271739) (H* 0.152174) (!H* 0.358696) (L+H* 0.184783) (L* 0.0326087) (L*+H 0) !H*))
192
+ ((syl_out is 2)
193
+ (((NONE 0.111111) (H* 0.361111) (!H* 0.319444) (L+H* 0.138889) (L* 0.0555556) (L*+H 0.0138889) H*))
194
+ ((syl_in is 4)
195
+ (((NONE 0.224) (H* 0.152) (!H* 0.328) (L+H* 0.24) (L* 0.056) (L*+H 0) !H*))
196
+ ((n.stress is 0)
197
+ ((syl_in is 3)
198
+ (((NONE 0.0833333) (H* 0.333333) (!H* 0.233333) (L+H* 0.216667) (L* 0.133333) (L*+H 0) H*))
199
+ (((NONE 0.283465) (H* 0.188976) (!H* 0.23622) (L+H* 0.204724) (L* 0.0708661) (L*+H 0.015748) NONE)))
200
+ (((NONE 0.305263) (H* 0.284211) (!H* 0.210526) (L+H* 0.178947) (L* 0.0210526) (L*+H 0) NONE))))))))))))))))))))))))))))))))))))))))
201
+ )
202
+
203
+ ; NON L-L L-H H-L
204
+ ; NONE13017 0 0 0 0 0 13017 [13017/13017] 100.000
205
+ ; H- 339 81 0 1 1 0 422 [81/422] 19.194
206
+ ; L- 223 52 0 5 0 0 280 [0/280] 0.000
207
+ ; L-L% 17 0 0 1057 96 0 1170 [1057/1170] 90.342
208
+ ; L-H% 16 0 0 457 139 0 612 [139/612] 22.712
209
+ ; H-L% 5 0 0 30 4 0 39 [0/39] 0.000
210
+ ; 13617 133 0 1550 240 0
211
+ ;total 15540 correct 14294.000 91.982%
212
+ (set! f2b_int_tone_cart_tree
213
+ '((lisp_syl_yn_question is 1)
214
+ (((H-H% 1.0) H-H%))
215
+ ((R:SylStructure.parent.gpos is cc)
216
+ (((NONE 0.996942) (H- 0.0030581) (L- 0) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
217
+ ((ssyl_in is 10)
218
+ (((NONE 0.989041) (H- 0.00273973) (L- 0) (L-L% 0.00273973) (L-H% 0.00547945) (H-L% 0) NONE))
219
+ ((R:SylStructure.parent.gpos is md)
220
+ (((NONE 0.986014) (H- 0) (L- 0) (L-L% 0.00699301) (L-H% 0.00699301) (H-L% 0) NONE))
221
+ ((p.old_syl_break is 4)
222
+ (((NONE 0.99462) (H- 0.00239091) (L- 0.00119546) (L-L% 0) (L-H% 0.00119546) (H-L% 0.000597729) NONE))
223
+ ((R:SylStructure.parent.gpos is det)
224
+ (((NONE 0.984635) (H- 0.00512164) (L- 0.00384123) (L-L% 0.00384123) (L-H% 0.00256082) (H-L% 0) NONE))
225
+ ((n.old_syl_break is 3)
226
+ (((NONE 0.981848) (H- 0.00495049) (L- 0.00330033) (L-L% 0.00660066) (L-H% 0.00330033) (H-L% 0) NONE))
227
+ ((n.old_syl_break is 4)
228
+ (((NONE 0.986982) (H- 0.000591716) (L- 0.0100592) (L-L% 0.00118343) (L-H% 0.00118343) (H-L% 0) NONE))
229
+ ((R:SylStructure.parent.gpos is in)
230
+ (((NONE 0.977865) (H- 0.00390625) (L- 0.00390625) (L-L% 0.0078125) (L-H% 0.00651042) (H-L% 0) NONE))
231
+ ((old_syl_break is 4)
232
+ ((R:SylStructure.parent.R:Word.n.gpos is 0)
233
+ (((NONE 0) (H- 0.00892857) (L- 0) (L-L% 0.982143) (L-H% 0.00892857) (H-L% 0) L-L%))
234
+ ((R:SylStructure.parent.R:Word.p.gpos is aux)
235
+ (((NONE 0) (H- 0) (L- 0) (L-L% 0.761905) (L-H% 0.238095) (H-L% 0) L-L%))
236
+ ((R:SylStructure.parent.R:Word.n.gpos is det)
237
+ (((NONE 0) (H- 0) (L- 0) (L-L% 0.652542) (L-H% 0.347458) (H-L% 0) L-L%))
238
+ ((ssyl_in is 4)
239
+ (((NONE 0) (H- 0) (L- 0) (L-L% 0.682243) (L-H% 0.313084) (H-L% 0.0046729) L-L%))
240
+ ((syl_in is 6)
241
+ (((NONE 0) (H- 0) (L- 0.00649351) (L-L% 0.688312) (L-H% 0.298701) (H-L% 0.00649351) L-L%))
242
+ ((R:SylStructure.parent.R:Word.n.gpos is aux)
243
+ (((NONE 0) (H- 0) (L- 0) (L-L% 0.464286) (L-H% 0.535714) (H-L% 0) L-H%))
244
+ ((syl_in is 5)
245
+ (((NONE 0) (H- 0) (L- 0) (L-L% 0.666667) (L-H% 0.322034) (H-L% 0.0112994) L-L%))
246
+ ((sub_phrases is 2)
247
+ (((NONE 0) (H- 0) (L- 0) (L-L% 0.696429) (L-H% 0.267857) (H-L% 0.0357143) L-L%))
248
+ ((R:SylStructure.parent.R:Word.p.gpos is det)
249
+ (((NONE 0) (H- 0) (L- 0) (L-L% 0.628866) (L-H% 0.350515) (H-L% 0.0206186) L-L%))
250
+ ((sub_phrases is 0)
251
+ ((R:SylStructure.parent.R:Word.n.gpos is in)
252
+ ((n.old_syl_break is 0)
253
+ (((NONE 0) (H- 0) (L- 0) (L-L% 0.68254) (L-H% 0.31746) (H-L% 0) L-L%))
254
+ (((NONE 0) (H- 0.0147059) (L- 0) (L-L% 0.338235) (L-H% 0.632353) (H-L% 0.0147059) L-H%)))
255
+ ((n.stress is 0)
256
+ (((NONE 0) (H- 0) (L- 0.0108303) (L-L% 0.599278) (L-H% 0.32491) (H-L% 0.064982) L-L%))
257
+ (((NONE 0) (H- 0) (L- 0) (L-L% 0.386364) (L-H% 0.579545) (H-L% 0.0340909) L-H%))))
258
+ (((NONE 0) (H- 0) (L- 0.00456621) (L-L% 0.652968) (L-H% 0.324201) (H-L% 0.0182648) L-L%))))))))))))
259
+ ((R:SylStructure.parent.gpos is pps)
260
+ (((NONE 0.988764) (H- 0.011236) (L- 0) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
261
+ ((syl_in is 0)
262
+ (((NONE 0.984848) (H- 0.0126263) (L- 0.00252525) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
263
+ ((R:SylStructure.parent.gpos is content)
264
+ ((R:SylStructure.parent.R:Word.nn.gpos is 0)
265
+ (((NONE 0.967914) (H- 0.0106952) (L- 0.0213904) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
266
+ ((pp.old_syl_break is 4)
267
+ (((NONE 0.972315) (H- 0.0232558) (L- 0.00442968) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
268
+ ((syl_in is 1)
269
+ (((NONE 0.951163) (H- 0.0372093) (L- 0.0116279) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
270
+ ((nn.old_syl_break is 4)
271
+ (((NONE 0.956244) (H- 0.0127621) (L- 0.0291705) (L-L% 0) (L-H% 0) (H-L% 0.00182315) NONE))
272
+ ((R:SylStructure.parent.R:Word.nn.gpos is in)
273
+ (((NONE 0.941919) (H- 0.0378788) (L- 0.020202) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
274
+ ((R:SylStructure.parent.R:Word.p.gpos is cc)
275
+ (((NONE 0.919643) (H- 0.0714286) (L- 0.00892857) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
276
+ ((nn.old_syl_break is 3)
277
+ (((NONE 0.927273) (H- 0.0472727) (L- 0.0254545) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
278
+ ((R:SylStructure.parent.R:Word.nn.gpos is cc)
279
+ (((NONE 0.921569) (H- 0.0588235) (L- 0.0196078) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
280
+ ((ssyl_in is 0)
281
+ (((NONE 0.911591) (H- 0.0825147) (L- 0.00589391) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
282
+ ((R:SylStructure.parent.R:Word.nn.gpos is to)
283
+ (((NONE 0.912281) (H- 0.0350877) (L- 0.0526316) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
284
+ ((R:SylStructure.parent.R:Word.pp.gpos is to)
285
+ (((NONE 0.894737) (H- 0.0526316) (L- 0.0526316) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
286
+ ((R:SylStructure.parent.R:Word.p.gpos is in)
287
+ (((NONE 0.888554) (H- 0.0662651) (L- 0.0451807) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
288
+ ((R:SylStructure.parent.R:Word.pp.gpos is in)
289
+ (((NONE 0.875817) (H- 0.0718954) (L- 0.0522876) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
290
+ ((syl_in is 2)
291
+ (((NONE 0.869942) (H- 0.0867052) (L- 0.0433526) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
292
+ ((R:SylStructure.parent.R:Word.nn.gpos is aux)
293
+ (((NONE 0.854839) (H- 0.0967742) (L- 0.0483871) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
294
+ ((sub_phrases is 1)
295
+ (((NONE 0.836538) (H- 0.0721154) (L- 0.0913462) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
296
+ ((R:SylStructure.parent.R:Word.pp.gpos is det)
297
+ (((NONE 0.832402) (H- 0.0949721) (L- 0.0726257) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
298
+ ((ssyl_in is 4)
299
+ (((NONE 0.793103) (H- 0.103448) (L- 0.103448) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
300
+ ((n.old_syl_break is 0)
301
+ (((NONE 0.850816) (H- 0.0839161) (L- 0.0652681) (L-L% 0) (L-H% 0) (H-L% 0) NONE))
302
+ ((R:SylStructure.parent.R:Word.n.gpos is content)
303
+ (((NONE 0.889447) (H- 0.0753769) (L- 0.0251256) (L-L% 0) (L-H% 0) (H-L% 0.0100503) NONE))
304
+ ((old_syl_break is 3)
305
+ (((NONE 0) (H- 0.609023) (L- 0.390977) (L-L% 0) (L-H% 0) (H-L% 0) H-))
306
+ (((NONE 1) (H- 0) (L- 0) (L-L% 0) (L-H% 0) (H-L% 0) NONE)))))))))))))))))))))))
307
+ (((NONE 0.978947) (H- 0.0131579) (L- 0.00789474) (L-L% 0) (L-H% 0) (H-L% 0) NONE)))))))))))))))
308
+
309
+ )
310
+
311
+ (defvar tobi_support_yn_questions t
312
+ "tobi_support_yn_questions
313
+ If set a crude final rise will be added at utterance that are judged
314
+ to be yesy/no questions. Namely ending in a ? and not starting with
315
+ a wh-for word.")
316
+
317
+ (define (first_word syl)
318
+ (let ((w (item.relation.parent syl 'SylStructure)))
319
+ (item.relation.first w 'Word)))
320
+
321
+ (define (syl_yn_question syl)
322
+ "(syl_yn_question utt syl)
323
+ Return 1 if this is the last syllable in a yes-no question. Basically
324
+ if it ends in question mark and doesn't start with a wh-woerd. This
325
+ isn't right but it depends on how much you want rising intonation."
326
+ (if (and
327
+ tobi_support_yn_questions
328
+ (member_string (item.feat syl "syl_break") '("4" "3"))
329
+ (not (member_string
330
+ (downcase (item.name (first_word syl)))
331
+ '("how" "why" "which" "who" "what" "where" "when")))
332
+ (string-matches
333
+ (item.feat syl "R:SylStructure.parent.R:Token.parent.punc")
334
+ ".*\\?.*"))
335
+ "1"
336
+ "0"))
337
+
338
+ (provide 'tobi)
CosyVoice-ttsfrd/resource/festival/tobi_rules.scm ADDED
@@ -0,0 +1,1002 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;; Authors: Robert A. J. Clark and Alan W Black
34
+ ;;; Modifications and Checking:
35
+ ;;; Gregor Moehler ([email protected])
36
+ ;;; Matthew Stone ([email protected])
37
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
38
+ ;;;
39
+ ;;; Generate F0 points from tobi labels using rules given in:
40
+ ;;; Jilka, Moehler & Dogil (forthcomming in Speech Communications)
41
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
42
+ ;;;
43
+ ;;; *** Converted to new Relation architecture -- but not checked yet -- awb
44
+ ;;; -> crude (beta) checking: gm in Dec. 98
45
+ ;;;
46
+ ;;; -> fixed TAKEOVER bug that used time value
47
+ ;;; as pitch target (!) - MDS 1/02
48
+ ;;; -> hacked around bunches of target overlap problems - MDS 1/02
49
+ ;;; -> added primitive pitch range controls
50
+ ;;;
51
+ ;;; Known problems and bugs:
52
+ ;;; Can't currently use voicing intervals which cross syllable boundaries,
53
+ ;;; so pre/post-nuclear tones are currently places 0.2s before/after the
54
+ ;;; nuclear tone even if no voicing occurs. Failing this they default a
55
+ ;;; percentage of the voicing for that syllable.
56
+ ;;;
57
+ ;;; Don't know about target points ahead of the current syllable.
58
+ ;;; (As you need to know what comes before them to calculate them)
59
+ ;;; So: post accent tones are placed 0.2 ahead if following syllable exists
60
+ ;;; ends before 0.2 from starred target and is not accented
61
+ ;;; The H-target of the H+!H* is 0.2 sec instead of 0.15 sec before
62
+ ;;; starred tone.
63
+ ;;;
64
+ ;;; Multi-utterance input has not been tested.
65
+ ;;;
66
+ ;;; !H- does not generate any targets
67
+ ;;;
68
+ ;;; Unfortunaltely some other modules may decide to put pauses in the
69
+ ;;; middle of a phrase
70
+ ;;;
71
+ ;;; valleys are not tested yet
72
+ ;;;
73
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
74
+ ;;;
75
+ ;;; To use this in a voice
76
+ ;;; (require 'tobi_rules)
77
+ ;;; And in the voice call
78
+ ;;; (setup_tobi_f0_method)
79
+ ;;; Set the following for your speaker's F0 range
80
+ ;;; (Parameter.set 'Default_Topline 146)
81
+ ;;; (Parameter.set 'Default_Start_Baseline 61)
82
+ ;;; (Parameter.set 'Valley_Dip 75)
83
+
84
+ ;; level of debug printout
85
+ (set! printdebug 0)
86
+
87
+ (define (setup_tobi_f0_method)
88
+ "(setup_tobi_f0_method)
89
+ Set up parameters for current voice to use the implementaion
90
+ of ToBI labels to F0 targets by rule."
91
+ (Parameter.set 'Int_Method Intonation_Tree)
92
+ (Parameter.set 'Int_Target_Method Int_Targets_General)
93
+ (set! int_accent_cart_tree no_int_cart_tree) ; NONE always
94
+ (set! int_tone_cart_tree no_int_cart_tree) ; NONE always
95
+ (set! int_general_params
96
+ (list
97
+ (list 'targ_func tobi_f0_targets))) ; we will return a list of f0 targets here
98
+
99
+ (Parameter.set 'Phrase_Method 'cart_tree)
100
+ (set! phrase_cart_tree tobi_label_phrase_cart_tree) ; redefines the phrasebreak tree
101
+ t)
102
+
103
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
104
+ ;;;;;;
105
+ ;;;;;; Define and set the new f0 rules
106
+ ;;;;;;
107
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
108
+
109
+ ;;; Set global parameters
110
+ ;;; You may want to reset these for different speakers
111
+
112
+ (Parameter.set 'Default_Topline 146) ;146
113
+ (Parameter.set 'Default_Start_Baseline 61) ;61
114
+ (Parameter.set 'Current_Topline (Parameter.get 'Default_Topline))
115
+ (Parameter.set 'Current_Start_Baseline (Parameter.get 'Default_Start_Baseline))
116
+ (Parameter.set 'Current_End_Baseline (Parameter.get 'Current_Start_Baseline))
117
+ (Parameter.set 'Downstep_Factor 0.70)
118
+ (Parameter.set 'Valley_Dip 75)
119
+ ;;; function to add target points on a given syllable and fill in
120
+ ;;; targets where necessary
121
+
122
+ (define (tobi_f0_targets utt syl)
123
+ "(tobi_f0_targets UTT ITEM)
124
+ Returns a list of targets for the given syllable."
125
+ (if (and (>= printdebug 1)
126
+ (not(equal? 0 (item.feat syl "R:Intonation.daughter1.name"))))
127
+ (format t "### %l (%.2f %.2f) %l ptarg: %l ###\n" (item.name syl)
128
+ (item.feat syl "syllable_start")(item.feat syl "syllable_end")
129
+ (item.feat syl "R:Intonation.daughter1.name") (ttt_last_target_time syl)))
130
+
131
+ ;; only continue if there is a Word related to this syllable
132
+ ;; I know there always should be, but there might be a bug elsewhere
133
+ (cond
134
+ ((not(equal? 0 (item.feat syl "R:SylStructure.parent.name")))
135
+
136
+ ; get current label. This assumes that there is only one accent and
137
+ ; one endtone on a syllable. Although there can be one of each.
138
+ (let ((voicing (ttt_get_voice_times syl)) ; voicing interval
139
+ (pvoicing (ttt_get_voice_times ; previous voicing
140
+ (item.relation.prev syl 'Syllable)))
141
+ (nvoicing (ttt_get_voice_times ; next voicing
142
+ (item.relation.next syl 'Syllable))))
143
+
144
+ ; if first syl of phrase set Phrase_Start and Phrase_End parameters
145
+ ; and reset downstep (currently does so on big and little breaks.)
146
+ ; only assignes Default values at this stage
147
+ ; maybe trained from CART later - first steps now - MDS
148
+ ; following Moehler and Mayer, SSW 2001
149
+ (if (eq 0 (item.feat syl 'syl_in)) ;; GM maybe something better needed here?
150
+ (progn
151
+ (Parameter.set 'Phrase_Start (item.feat syl 'R:SylStructure.parent.R:Phrase.last.word_start))
152
+ (Parameter.set 'Phrase_End (item.feat syl 'R:SylStructure.parent.R:Phrase.last.word_end))
153
+ (Parameter.set 'Current_Topline
154
+ (/ (* (wagon syl ttt_topline_tree)
155
+ (Parameter.get 'Default_Topline)) 100))
156
+ (Parameter.set 'Current_Start_Baseline
157
+ (/ (* (wagon syl ttt_baseline_tree)
158
+ (Parameter.get 'Default_Start_Baseline)) 100))
159
+ (Parameter.set 'Current_End_Baseline
160
+ (Parameter.get 'Current_Start_Baseline))
161
+ (if (>= printdebug 3)
162
+ (begin
163
+ (print (format nil "new range: %f %f %f"
164
+ (Parameter.get 'Current_Topline)
165
+ (Parameter.get 'Current_Start_Baseline)
166
+ (Parameter.get 'Current_End_Baseline) )))) ))
167
+
168
+ ; do stuff (should go only if there is an accent/boundary?)
169
+ (let ((new_targets
170
+ (ttt_to_targets syl (wagon syl ttt_starttone_tree)
171
+ voicing
172
+ pvoicing
173
+ nvoicing
174
+ 'Starttones)))
175
+
176
+ (set! new_targets (append new_targets
177
+ (ttt_to_targets syl (wagon syl ttt_accent_tree)
178
+ voicing
179
+ pvoicing
180
+ nvoicing
181
+ 'Accents)))
182
+
183
+ (set! new_targets (append new_targets
184
+ (ttt_to_targets syl (wagon syl ttt_endtone_tree)
185
+ voicing
186
+ pvoicing
187
+ nvoicing
188
+ 'Endtones)))
189
+
190
+ (if (and(not(equal? new_targets nil))
191
+ (>= printdebug 2))
192
+ (begin
193
+ (format t ">> Targets: %l\n" new_targets)
194
+ (format t ">> LastTarget: %l\n" (last new_targets))
195
+ ))
196
+
197
+ new_targets)))))
198
+
199
+
200
+ ;;; CART tree to specify no accents
201
+
202
+ (set! no_int_cart_tree
203
+ '
204
+ ((NONE)))
205
+
206
+ ;;;
207
+ ;;; Relate phrasing to boundary tones.
208
+ ;;; Added downstepped tones - MDS
209
+
210
+ (set! tobi_label_phrase_cart_tree
211
+ '
212
+ ((tone in ("L-" "H-" "!H-"))
213
+ ((B))
214
+ ((tone in ("H-H%" "H-L%" "!H-L%" "L-L%" "L-H%"))
215
+ ((BB))
216
+ ((NB)))))
217
+
218
+ ;;;
219
+ ;;; The other functions
220
+ ;;;
221
+
222
+ ;;; process a list of relative targets and convert to actual targets
223
+
224
+ (define (ttt_to_targets syl rlist voicing pvoicing nvoicing type)
225
+ "Takes a list of target sets and returns a list of targets."
226
+ (if (or (and (>= printdebug 2)
227
+ rlist (atom (caar rlist))
228
+ (not (equal? 'NONE (caar rlist))) (not (equal? '(NONE) (caar rlist))))
229
+ (>= printdebug 3))
230
+ (begin (print "Entering ttt_to_targets with:")
231
+ (print (format nil "rlist: %l vc: %l pvc: %l nvc: %l type: %s" rlist voicing pvoicing nvoicing type))))
232
+ (cond
233
+ ;; nowt
234
+ ((eq (length rlist) 0) ())
235
+ ;; a single target set
236
+ ((atom (car (car rlist)))
237
+ (cond
238
+ ((eq type 'Accents)
239
+ (ttt_accent_set_to_targets syl rlist voicing pvoicing nvoicing))
240
+ ((eq type 'Starttones)
241
+ (ttt_bound_set_to_targets syl rlist voicing pvoicing))
242
+ ((eq type 'Endtones)
243
+ (ttt_bound_set_to_targets syl rlist voicing pvoicing))
244
+ (t (error "unknown target set encountered in ttt_to_targets"))))
245
+ ;; list of target sets
246
+ ((atom (car (car (car rlist))))
247
+ (append (ttt_to_targets syl (cdr rlist) voicing pvoicing nvoicing type)
248
+ (ttt_to_targets syl (car rlist) voicing pvoicing nvoicing type)))
249
+ ;; error
250
+ (t (error "something strange has happened in ttt_to_targets"))))
251
+
252
+
253
+ ;; process a starttone/endtone target set.
254
+
255
+ (define (ttt_bound_set_to_targets syl tset voicing pvoicing)
256
+ "takes a start/endtone target set and returns a list of target points."
257
+ (if (>= printdebug 3) (begin
258
+ (print "Entering ttt_bound_set_to_targets with:")
259
+ (pprintf (format nil "tset: %l vc: %l pvc: %l" tset voicing pvoicing))))
260
+ (cond
261
+ ;; usually target given is NONE. (also ignore unknown!)
262
+ ((or (eq (car (car tset)) 'NONE)
263
+ (eq (car (car tset)) 'UNKNOWN))
264
+ nil)
265
+ ;; a pair of target pairs
266
+ ((eq (length tset) 2)
267
+ (list (ttt_get_target (car tset) voicing)
268
+ (ttt_get_target (car (cdr tset)) voicing)))
269
+ ;; single target pair
270
+ ((eq (length tset) 1)
271
+ (cond
272
+ ;; an actual target pair
273
+ ((not (null (cdr (car tset))))
274
+ (list (ttt_get_target (car tset) voicing)))
275
+ ;; a TAKEOVER marker
276
+ ((eq (car (car tset)) 'TAKEOVER)
277
+ (list (list (ttt_interval_percent voicing 0)
278
+ (ttt_last_target_value syl))))
279
+ (t (error "unknown target pair in ttt_bound_set_to_targets"))))
280
+ (t (error "unknown target set type in ttt_bound_set_to_targets"))))
281
+
282
+
283
+ ;; process an accent target set.
284
+
285
+ (define (ttt_accent_set_to_targets syl tset voicing pvoicing nvoicing)
286
+ "takes a accent target set and returns a list of target points."
287
+ (if (>= printdebug 3) (begin
288
+ (print "Entering ttt_accent_set_to_targets with:")
289
+ (pprintf (format nil "tset: %l vc: %l pvc: %l nvc: %l" tset voicing pvoicing nvoicing))))
290
+ (cond
291
+ ;; single target in set
292
+ ((null (cdr tset))
293
+ (cond
294
+ ; target given is NONE.
295
+ ((or (eq (car (car tset)) 'NONE)
296
+ (eq (car (car tset)) 'UNKNOWN)) nil)
297
+ ; V1 marker
298
+ ((eq (car (car tset)) 'V1)
299
+ (let ((target_time (+ (/ (- (next_accent_start syl)
300
+ (ttt_last_target_time syl))
301
+ 2.0)
302
+ (ttt_last_target_time syl))))
303
+ (list (list target_time (ttt_accent_pitch (Parameter.get 'Valley_Dip) target_time)))))
304
+ ; V2 marker
305
+ ((eq (car (car tset)) 'V2)
306
+ (let ((target_time (+ (ttt_last_target_time syl) 0.25)))
307
+ (list (list target_time (ttt_accent_pitch (Parameter.get 'Valley_Dip) target_time)))))
308
+ ; V3 marker
309
+ ((eq (car (car tset)) 'V3)
310
+ (let ((target_time (- (next_accent_start syl) 0.25)))
311
+ (list (list target_time (ttt_accent_pitch (Parameter.get 'Valley_Dip) target_time)))))
312
+ ; single target pair
313
+ (t (list (ttt_get_target (car tset) voicing)))))
314
+ ;; a pair of targets
315
+ ((length tset 2)
316
+ (cond
317
+ ;; a *ed tone with PRE type tone (as in L+H*)
318
+ ((eq (car (car tset)) 'PRE)
319
+ (let ((star_target (ttt_get_target (car (cdr tset)) voicing))
320
+ (last_target (parse-number(ttt_last_target_time syl))))
321
+ (cond
322
+ ; normal 0.2s case (currently doesn't check for voicing)
323
+ ((and (eqv? 0 (ip_initial syl))
324
+ (> (- (car star_target) 0.2) last_target))
325
+ (list (list (- (car star_target) 0.2)
326
+ (ttt_accent_pitch (car (cdr (car tset)))
327
+ (- (car star_target) 0.2))) ; the time
328
+ star_target))
329
+
330
+ ; 90% prev voiced if not before last target - Added back in MDS,
331
+ ; with parse-number added and new check for ip_initial
332
+ ((and (eqv? 0 (ip_initial syl))
333
+ (> (parse-number (ttt_interval_percent pvoicing 90))
334
+ (parse-number (ttt_last_target_time syl))))
335
+ (list (list (ttt_interval_percent pvoicing 90)
336
+ (ttt_accent_pitch (car (cdr (car tset)))
337
+ (ttt_interval_percent pvoicing 90)))
338
+ star_target))
339
+
340
+ ; otherwise (UNTESTED) [NOTE: Voicing for this syllable only]
341
+ (t
342
+ (list (list (ttt_interval_percent voicing 20)
343
+ (ttt_accent_pitch (car (cdr (car tset)))
344
+ (ttt_interval_percent voicing 20)))
345
+ star_target)))))
346
+ ; a *ed tone with POST type tone (as L*+H)
347
+ ((eq (car(car(cdr tset))) 'POST)
348
+ (let ((star_target (ttt_get_target (car tset) voicing))
349
+ (next_target nil ) ; interesting problem
350
+ (next_syl (item.next syl)))
351
+
352
+ (cond
353
+ ; normal 0.2s case (UNTESTED)
354
+ ((and (not (equal? next_syl nil))
355
+ (eq 0 (item.feat next_syl "accented")))
356
+ (cond
357
+ ((< (+ (car star_target) 0.2) (item.feat next_syl "syllable_end"))
358
+ (list star_target
359
+ (list (+ (car star_target) 0.2)
360
+ (ttt_accent_pitch (car (cdr (car (cdr tset))))
361
+ (+ (car star_target) 0.2) ))))
362
+ (t
363
+
364
+ (list star_target
365
+ (list (ttt_interval_percent nvoicing 90)
366
+ (ttt_accent_pitch (car (cdr (car (cdr tset))))
367
+ (ttt_interval_percent nvoicing 90) ))))))
368
+
369
+ ; 20% next voiced (BUG: Can't do this as the next target hasn't been
370
+ ; calculated yet!)
371
+ (nil nil)
372
+ ;otherwise (UNTESTED)
373
+ (t (list star_target
374
+ (list (ttt_interval_percent voicing 90)
375
+ (ttt_accent_pitch (car (cdr (car (cdr tset))))
376
+ (ttt_interval_percent voicing 90) )))))))
377
+
378
+ (t
379
+ ;; This case didn't use to happen, but now must
380
+ ;; to avoid +H's clobbering endtones - MDS's hack.
381
+ (list (ttt_get_target (car tset) voicing)
382
+ (ttt_get_target (cadr tset) voicing)))))
383
+
384
+
385
+ ;; something else...
386
+ (t (error (format nil "unknown accent set in ttt_accent_set_to_targets: %l" tset)))))
387
+
388
+
389
+
390
+ (define (ttt_get_target pair voicing)
391
+ "Returns actual target pair, usually for a stared tone."
392
+ (if (>= printdebug 4) (begin
393
+ (print "Entering ttt_get_target with:")
394
+ (pprintf pair) (pprintf voicing)))
395
+ (list (ttt_interval_percent voicing (car pair))
396
+ (ttt_accent_pitch (car (cdr pair))
397
+ (ttt_interval_percent voicing (car pair)))))
398
+
399
+ (define (ttt_accent_pitch value time)
400
+ "Converts a accent pitch entry to a pitch value."
401
+ (if (>= printdebug 4) (begin
402
+ (print "Entering ttt_accent_pitch with:")
403
+ (pprintf value)))
404
+ (cond
405
+ ;; a real value
406
+ ((number? value)
407
+ (ttt_interval_percent (list (ttt_get_current_baseline time)
408
+ (Parameter.get 'Current_Topline))
409
+ value))
410
+ ;; Downstep then Topline
411
+ ((eq value 'DHIGH)
412
+ (progn
413
+ (Parameter.set 'Current_Topline (+ (ttt_get_current_baseline time)
414
+ (* (Parameter.get 'Downstep_Factor)
415
+ (- (Parameter.get 'Current_Topline)
416
+ (ttt_get_current_baseline time)))))
417
+ (ttt_interval_percent (list (ttt_get_current_baseline time)
418
+ (Parameter.get 'Current_Topline))
419
+ 100)))
420
+
421
+ ;; Unknown
422
+ (t (error "Unknown accent pitch value encountered"))))
423
+
424
+
425
+ (define (ttt_get_current_baseline v)
426
+ "Returns the current declined baseline at time v."
427
+ (if (>= printdebug 4) (begin
428
+ (print "Entering ttt_get_current_baseline with:")
429
+ (pprintf v)))
430
+ (let ((h (Parameter.get 'Current_Start_Baseline))
431
+ (l (Parameter.get 'Current_End_Baseline))
432
+ (e (Parameter.get 'Phrase_End))
433
+ (s (Parameter.get 'Phrase_Start)))
434
+ (- h (* (/ (- h l) (- e s)) (- v s)))))
435
+
436
+ ;;; find the time n% through an inteval
437
+
438
+ (define (ttt_interval_percent pair percent)
439
+ "Returns the time that is percent percent thought the pair."
440
+ (if (>= printdebug 4) (begin
441
+ (print "Entering ttt_interval_percent with:")
442
+ (pprintf (format nil "%l, %l" pair percent))))
443
+ (cond
444
+ ; no pair given: just return nil
445
+ ((null pair) nil)
446
+ ; otherwise do the calculation
447
+ (t (let ((start (car pair))
448
+ (end (car(cdr pair))))
449
+ (+ start (* (- end start) (/ percent 100)))))))
450
+
451
+
452
+ ;;; Getting start and end voicing times in a syllable
453
+
454
+ (define (ttt_get_voice_times syl_item)
455
+ "Returns a pair of start time of first voiced phone in syllable and
456
+ end of last voiced phone in syllable, or nil if syllable is nil"
457
+ (cond
458
+ ((null syl_item) nil)
459
+ (t (let ((segs (item.relation.daughters syl_item "SylStructure")))
460
+ (list
461
+ (item.feat (ttt_first_voiced segs) "segment_start")
462
+ (item.feat (ttt_first_voiced (reverse segs)) "end"))))))
463
+
464
+ (define (ttt_first_voiced segs)
465
+ "Returns first segment that is voiced (vowel or voiced consonant)
466
+ returns last segment if all are unvoiced."
467
+ (cond
468
+ ((null (cdr segs))
469
+ (car segs)) ;; last possibility
470
+ ((equal? "+" (item.feat (car segs) "ph_vc"))
471
+ (car segs))
472
+ ((equal? "+" (item.feat (car segs) "ph_cvox"))
473
+ (car segs))
474
+ (t
475
+ (ttt_first_voiced (cdr segs)))))
476
+
477
+ ;;; ttt_last_target has bifurcated into
478
+ ;;; ttt_last_target_time and
479
+ ;;; ttt_last_target_value
480
+ ;;; to fix a place where f0 was set to last target time!
481
+ ;;; - MDS
482
+
483
+ (define (ttt_last_target_time syl)
484
+ "Returns the end of the most recent previous target
485
+ in the utterance or nil if there is not one present
486
+ "
487
+ (if (>= printdebug 3)
488
+ (begin (print "Entering ttt_last_target_time")
489
+ (print syl))
490
+ )
491
+ (let ((target (ttt_last_target syl)))
492
+ (if (null? target)
493
+ nil
494
+ (item.feat target "R:Target.daughter1.pos"))))
495
+
496
+ (define (ttt_last_target_value syl)
497
+ "Returns the pitch of the most recent previous target
498
+ in the utterance or nil if there is not one present
499
+ "
500
+ (if (>= printdebug 3)
501
+ (begin (print "Entering ttt_last_target_time")
502
+ (print syl))
503
+ )
504
+ (let ((target (ttt_last_target syl)))
505
+ (if (null? target)
506
+ nil
507
+ (item.feat target "R:Target.daughter1.f0"))))
508
+
509
+ ;; Changed to scan through segments in the segment relation,
510
+ ;; to catch (notional) targets on pauses. - MDS
511
+ ;;
512
+ ;;; associated segments are:
513
+ ;;; - the segments in the word
514
+ ;;; - subsequent segments not in the syllable structure
515
+ ;;; and on the first word, preceding segments
516
+ ;;; not in the syllable structure
517
+
518
+ (define (ttt_collect_following seg accum)
519
+ (if (or (null? seg)
520
+ (not (null? (item.relation seg 'SylStructure))))
521
+ accum
522
+ (ttt_collect_following (item.next seg)
523
+ (cons seg accum))))
524
+
525
+
526
+ (define (ttt_last_target syl)
527
+ "Returns the most recent previous target
528
+ in the utterance or nil if there is not one present
529
+ "
530
+ (if (>= printdebug 3)
531
+ (begin (print "Entering ttt_last_target")
532
+ (print syl))
533
+ )
534
+ (let ((prev_syl (item.relation.prev syl 'Syllable)))
535
+ (cond
536
+ ; ((symbol-bound? 'new_targets) (last (caar new_targets)))
537
+ ((null prev_syl) nil)
538
+ ((ttt_last_target_segs
539
+ (ttt_collect_following
540
+ (item.relation.next
541
+ (item.relation.daughtern prev_syl "SylStructure")
542
+ "Segment")
543
+ (reverse (item.relation.daughters prev_syl "SylStructure")))))
544
+ ;list of segments of prev. syllable
545
+ ;in reverse order, with pauses
546
+ ;prepended.
547
+ (t (ttt_last_target prev_syl)))))
548
+
549
+ (define (ttt_last_target_segs segs)
550
+ "Returns the first target no earlier than seg
551
+ or nil if there is not one
552
+ "
553
+ (if (>= printdebug 4)
554
+ (begin (print "Entering ttt_last_target_segs with:")
555
+ (pprintf (format nil "%l" segs))
556
+ ))
557
+ (cond
558
+ ((null segs) nil)
559
+ ((and (> (parse-number
560
+ (item.feat (car segs) "R:Target.daughter1.f0")) 0)
561
+ (eq 0 (item.feat (car segs) "R:SylStructure.parent.lisp_lh_condition"))
562
+ (eq 0 (item.feat (car segs) "R:SylStructure.parent.lisp_hl_condition"))
563
+ (eq 0 (item.feat (car segs) "R:SylStructure.parent.lisp_valley_condition")))
564
+ (car segs))
565
+
566
+ (t (ttt_last_target_segs (cdr segs)))))
567
+
568
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
569
+ ;;;;;;
570
+ ;;;;;; CART TREES (ttt - tobi to target)
571
+ ;;;;;;
572
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
573
+
574
+ ;;;
575
+ ;;; Return a list of target lists. A target list comprises of a list
576
+ ;;; of related targets (ie for the L and H in L+H*), just to confuse
577
+ ;;; matters each target is also a list! (pos pitch)
578
+ ;;;
579
+
580
+
581
+ (set! ttt_endtone_tree ; BUG: does it check the current syl for last accent?
582
+ '
583
+ ((tobi_endtone is NONE) ; ususally none
584
+ ((((NONE))))
585
+ ((tobi_endtone is "H-H%") ; H-H%
586
+ ((((100 120))))
587
+ ((tobi_endtone is "L-L%") ; L-L%
588
+ ((((100 -20))))
589
+ ((tobi_endtone is "L-H%") ; L-H%
590
+ ((lisp_last_accent > 2)
591
+ ((lisp_last_accent_type is "L*")
592
+ ((((0 25) (100 40)))) ; paper says 80 but AWB had 40
593
+ ((((0 0) (100 40)))))
594
+ ((lisp_last_accent_type is "L*")
595
+ ((((100 40))))
596
+ ((((50 0) (100 40))))))
597
+ ((tobi_endtone is "H-L%") ; H-L%
598
+ ((lisp_last_accent_type is "L*")
599
+ ((tobi_accent is"L*")
600
+ ((((50 100) (100 100))))
601
+ ((((0 100) (100 100)))))
602
+ ((((100 100)))))
603
+ ((tobi_endtone is "!H-L%") ; !H-L%
604
+ ((lisp_last_accent_type is "L*")
605
+ ((tobi_accent is"L*")
606
+ ((((50 DHIGH) (100 100))))
607
+ ((((0 DHIGH) (100 100)))))
608
+ ((((100 DHIGH)))))
609
+ ((tobi_endtone is "H-")
610
+ ((((100 100))))
611
+ ((tobi_endtone is "!H-")
612
+ ((((100 DHIGH))))
613
+ ((tobi_endtone is "L-")
614
+ ((((100 0))))
615
+ ((((UNKNOWN))))))))))))))
616
+
617
+ (set! ttt_starttone_tree
618
+ '
619
+ ((lisp_ip_initial = 1)
620
+ ((tobi_endtone is "%H")
621
+ ((((0 100))))
622
+ ((p.tobi_endtone in ("H-" "!H-" "L-"))
623
+ ((((TAKEOVER)))) ; takeover case
624
+ ((tobi_accent is NONE)
625
+ ((lisp_next_accent > 2) ; default cases (dep. on whether next target is low)
626
+ ((lisp_next_accent_type in ("L*" "L*+H" "L*+!H" "L+H*" "L+!H*" "L-" "L-H%" "L-L%"))
627
+ ((((0 50)(100 25))))
628
+ ((((0 50)(100 75)))))
629
+ ((lisp_next_accent_type in ("L*" "L*+H" "L*+!H" "L+H*" "L+!H*" "L-" "L-H%" "L-L%"))
630
+ ((((0 30))))
631
+ ((((0 70))))))
632
+ ((tobi_accent in ("L*" "L*+H" "L*+!H" "L+H*" "L+!H*" "L-" "L-H%" "L-L%"))
633
+ ((((0 30))))
634
+ ((((0 70))))))))
635
+ ((((NONE)))))) ; otherwise (and usually) nothing.
636
+
637
+ ;; Redone after Jilka, Moehler and Dogil
638
+ ;; - But treating one-syllable-ip's like
639
+ ;; last-syllable-of-ip's in cases of
640
+ ;; two tone switches per syllable (e.g. H* L-H%).
641
+ ;; - And (hack) a 70% target for the initial
642
+ ;; H*'s of phrases when the next accent is L+H*
643
+ ;; - MDS
644
+
645
+ (set! ttt_accent_tree
646
+ '
647
+ ((tobi_accent is "H*" ) ; H*
648
+ ((lisp_ip_final = 1)
649
+ ((lisp_ip_one_syllable_case = 1)
650
+ ((((50 100))))
651
+ ((((25 100)))))
652
+ ((lisp_hstar_weak_target = 1)
653
+ ((((60 70))))
654
+ ((lisp_ip_initial = 1)
655
+ ((((85 100))))
656
+ ((((60 100)))))))
657
+
658
+ ((tobi_accent is "!H*" ) ; !H*
659
+ ((lisp_ip_final = 1)
660
+ ((lisp_ip_one_syllable_case = 1)
661
+ ((((50 DHIGH))))
662
+ ((((25 DHIGH)))))
663
+ ((lisp_ip_initial = 1)
664
+ ((((85 DHIGH))))
665
+ ((((60 DHIGH))))))
666
+
667
+ ((tobi_accent is "L*" ) ; L*
668
+ ((lisp_ip_final = 1)
669
+ ((lisp_ip_one_syllable_case = 1)
670
+ ((((50 0))))
671
+ ((((25 0)))))
672
+ ((lisp_ip_initial = 1)
673
+ ((((85 0))))
674
+ ((((60 0))))))
675
+
676
+ ((tobi_accent is "L+H*") ; L+H*
677
+ ((lisp_ip_final = 1)
678
+ ((lisp_ip_one_syllable_case = 1)
679
+ ((((PRE 20) (50 100)))) ; JMD estimated 70
680
+ ((((PRE 20) (25 100)))))
681
+ ((lisp_ip_initial = 1)
682
+ ((((PRE 20) (90 100))))
683
+ ((((PRE 20) (75 100))))))
684
+
685
+ ((tobi_accent is "L+!H*") ; L+!H*
686
+ ((lisp_ip_final = 1)
687
+ ((lisp_ip_one_syllable_case = 1)
688
+ ((((PRE 20) (70 DHIGH))))
689
+ ((((PRE 20) (25 DHIGH)))))
690
+ ((lisp_ip_initial = 1)
691
+ ((((PRE 20) (90 DHIGH))))
692
+ ((((PRE 20) (75 DHIGH))))))
693
+
694
+ ((tobi_accent is "L*+H") ; L*+H
695
+ ((lisp_ip_final = 1)
696
+ ((lisp_ip_one_syllable_case = 1)
697
+ ((((35 0) (80 100)))) ; POST would clobber endtones
698
+ ((((15 0) (40 100))))) ; POST would clobber endtones - MDS
699
+ ((lisp_ip_initial = 1)
700
+ ((((55 0) (POST 100))))
701
+ ((((40 0) (POST 100))))))
702
+
703
+ ((tobi_accent is "L*+!H") ; L*+!H
704
+ ((lisp_ip_final = 1)
705
+ ((lisp_ip_one_syllable_case = 1)
706
+ ((((35 0) (80 DHIGH)))) ; POST would clobber endtones - MDS
707
+ ((((15 0) (40 DHIGH))))) ; POST would clobber endtones - MDS
708
+ ((lisp_ip_initial = 1)
709
+ ((((55 0) (POST DHIGH))))
710
+ ((((40 0) (POST DHIGH))))))
711
+
712
+ ((tobi_accent is "H+!H*") ; H+!H*
713
+ ((lisp_ip_final = 1)
714
+ ((lisp_ip_one_syllable_case = 1)
715
+ ((((PRE 143) (60 DHIGH)))) ; the 143 is a hack to level out the downstep
716
+ ((((PRE 143) (20 DHIGH)))))
717
+ ((lisp_ip_initial = 1)
718
+ ((((PRE 143) (90 DHIGH))))
719
+ ((((PRE 143) (60 DHIGH))))))
720
+
721
+ ((lisp_lh_condition = 1)
722
+ ((((100 75))))
723
+ ((lisp_lh_condition = 2)
724
+ ((((0 90))))
725
+ ((lisp_hl_condition = 1)
726
+ ((((100 25))))
727
+ ((lisp_valley_condition = 1)
728
+ ((((V1 85))))
729
+ ((lisp_valley_condition = 2)
730
+ ((((V2 70))))
731
+ ((lisp_valley_condition = 3)
732
+ ((((V3 70))))
733
+ ((tobi_accent is NONE) ; usually we find no accent
734
+ ((((NONE))))
735
+ ((((UNKNOWN)))))))))))))))))))) ; UNKNOWN TARGET FOUND
736
+
737
+ ;;; Cart tree to "predict" pitch range
738
+ ;;; Right now just accesses a feature
739
+ ;;; "register" following Moehler & Mayer 2001.
740
+ ;;; Register must be one of
741
+ ;;; H - primary high register (default): 133% lowest, 92% highest
742
+ ;;; H-H - expanded high register: 134% lowest, 100% highest
743
+ ;;; H-L - lowered high register: 128% lowest, 87% highest
744
+ ;;; L - primary low register: 100% lowest, 73% highest
745
+ ;;; L-L and HL-L - low compressed: 100% lowest, 66% highest
746
+ ;;; HL - expanded register: 100% lowest, 84% highest
747
+ ;;; HL-H - complete register: 100% lowest, 96% highest
748
+ ;;; For their speaker, ,BASELINE was 42% of PEAK
749
+
750
+ (set! ttt_topline_tree
751
+ '
752
+ ((R:SylStructure.parent.register is "H")
753
+ (92)
754
+ ((R:SylStructure.parent.register is "H-H")
755
+ (100)
756
+ ((R:SylStructure.parent.register is "H-L")
757
+ (87)
758
+ ((R:SylStructure.parent.register is "L")
759
+ (73)
760
+ ((R:SylStructure.parent.register is "L-L")
761
+ (66)
762
+ ((R:SylStructure.parent.register is "HL")
763
+ (84)
764
+ ((R:SylStructure.parent.register is "HL-H")
765
+ (96)
766
+ (92)))))))))
767
+
768
+ (set! ttt_baseline_tree
769
+ '
770
+ ((R:SylStructure.parent.register is "H")
771
+ (133)
772
+ ((R:SylStructure.parent.register is "H-H")
773
+ (134)
774
+ ((R:SylStructure.parent.register is "H-L")
775
+ (128)
776
+ ((R:SylStructure.parent.register is "L")
777
+ (100)
778
+ ((R:SylStructure.parent.register is "L-L")
779
+ (100)
780
+ ((R:SylStructure.parent.register is "HL")
781
+ (100)
782
+ ((R:SylStructure.parent.register is "HL-H")
783
+ (100)
784
+ (133)))))))))
785
+
786
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
787
+ ;;;;;;
788
+ ;;;;;; Lisp Feature functions.
789
+ ;;;;;;
790
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
791
+
792
+ (define (valley_condition syl)
793
+ "(valley_condition syl)
794
+ Function to determine if a lowered target between two high target points
795
+ is needed in this syllable.
796
+ Returns: 0 - no target required
797
+ 1 - the single target case
798
+ 2 - the first of the two target case
799
+ 3 - the second of the two target case
800
+ "
801
+ (if (>= printdebug 4)
802
+ (begin (print "Entering valley_condition")))
803
+ (cond
804
+ ((and (eq 0 (item.feat syl 'accented))
805
+ (string-matches (next_accent_type syl)
806
+ "\\(H\\*\\|H\\-\\|H\\-L\\%\\|H\\-H\\%\\|\\!H\\*\\|\\!H\\-\\|\\!H\\-L\\%\\|\\!H\\-H\\%\\)")
807
+ (string-matches (last_accent_type syl)
808
+ "\\(H\\*\\|L\\+H\\*\\|L\\*\\+H\\\\|\\!H\\*\\|L\\+\\!H\\*\\|L\\*\\+\\!H\\)"))
809
+ ;GM: excluded %H (returns nil for last target)
810
+ (let ((nas (next_accent_start syl))
811
+ (syls (item.feat syl 'syllable_start))
812
+ (syle (item.feat syl 'syllable_end))
813
+ (las (ttt_last_target_time syl)))
814
+ (if (>= printdebug 3)
815
+ (begin (print (format nil "nas: %l syls: %l syle %l las %l" nas syls syle las))))
816
+ (cond
817
+ ((and (< (- nas las) 0.5)
818
+ (> (- nas las) 0.25)
819
+ (< syls (+ (/ (- nas las) 2.0) (ttt_last_target_time syl)))
820
+ (> syle (+ (/ (- nas las) 2.0) (ttt_last_target_time syl)))) 1)
821
+ ((and (> (- nas las) 0.5)
822
+ (< syls (+ (ttt_last_target_time syl) 0.25))
823
+ (> syle (+ (ttt_last_target_time syl) 0.25))) 2)
824
+ ((and (> (- nas las) 0.5)
825
+ (< syls (- nas 0.25))
826
+ (> syle (- nas 0.25))) 3)
827
+ (t 0))))
828
+ (t 0)))
829
+
830
+
831
+
832
+ (define (lh_condition syl)
833
+ "(lh_condition syl)
834
+ Function to determine the need for extra target points between an L and an H
835
+ Returns: 1 - first extra target required
836
+ 2 - second extra target required
837
+ 0 - no target required.
838
+ "
839
+ (if (>= printdebug 4)
840
+ (begin (print "Entering LH_condition")))
841
+ (cond
842
+ ((and (eq 0 (item.feat syl 'accented))
843
+ (string-matches (last_accent_type syl) "\\(L\\*\\)")
844
+ (string-matches (next_accent_type syl)
845
+ "\\(H\\*\\|H\\-\\|H\\-L\\%\\|H\\-H\\%\\)"))
846
+ (cond
847
+ ((and (eq 1 (last_accent syl))
848
+ (< 2 (next_accent syl))) 1)
849
+ ((and (< 2 (last_accent syl))
850
+ (eq 1 (next_accent syl))) 2)
851
+ (t 0)))
852
+ (t 0)))
853
+
854
+ (define (hl_condition syl)
855
+ "(lh_condition syl)
856
+ Function to determine the need for extra target points between an H and an L
857
+ Returns: 1 - extra target required
858
+ 0 - no target required.
859
+ "
860
+ (if (>= printdebug 4)
861
+ (begin (print "Entering HL_condition")))
862
+ (cond
863
+ ((and (eq 0 (item.feat syl 'accented))
864
+ (string-matches (next_accent_type syl)
865
+ "\\(L\\*\\|L\\+H\\*\\|L\\+\\!H\\*\\|L\\*\\+H\\|L\\*\\+!H\\|L\\-\\|L\\-L\\%\\|L-H\\%\\)")
866
+ (string-matches (last_accent_type syl)
867
+ "\\(H\\*\\|L\\+H\\*\\|L\\*\\+H\\\\|\\!H\\*\\|L\\+\\!H\\*\\|L\\*\\+\\!H\\|\\%H\\)")
868
+ ;MDS: added !H's
869
+ (eq 1 (last_accent syl))
870
+
871
+ ;; fall faster! -MDS
872
+ (<= 2 (next_accent syl))) 1)
873
+ (t 0)))
874
+
875
+ (define (next_accent syl)
876
+ "(next_accent syl)
877
+ Wrapper for c++ func ff_next_accent.
878
+ Returns the number of the syllables to the next accent in the following format.
879
+ 0 - no next accent
880
+ 1 - next syllable
881
+ 2 - next next syllable
882
+ etc..."
883
+ (if (>= printdebug 4)
884
+ (begin (print "Entering next_accent")))
885
+ (cond
886
+ ((eq 0 (next_accent_type syl)) 0)
887
+ (t (+ (item.feat syl 'next_accent) 1))))
888
+
889
+ ;; Fixed bug that crashed complex phrase tones. - MDS
890
+ ;; Not sure how else to get a big number...
891
+ (define infinity (/ 1 0))
892
+
893
+ ;; Modified to include current accent as well -MDS
894
+
895
+ (define (last_accent syl)
896
+ "(last_accent syl)
897
+ Wrapper for c++ func ff_last_accent.
898
+ Returns the number of the syllables to the previous accent in the following format.
899
+ 0 - accent on current syllable
900
+ 1 - prev syllable
901
+ 2 - prev to prev syllable
902
+ etc...
903
+ infinity - no previous syllable"
904
+ (if (>= printdebug 4)
905
+ (begin (print "Entering last_accent")))
906
+ (cond
907
+ ((not (equal? "NONE" (item.feat syl 'tobi_accent))) 0)
908
+ ((equal? 0 (last_accent_type syl)) infinity)
909
+ (t (+ (item.feat syl 'last_accent) 1))))
910
+
911
+ (define (next_accent_type syl)
912
+ "(next_accent_type syl)
913
+ Returns the type of the next accent."
914
+ (cond
915
+ ((not (eq 0 (item.feat syl "n.R:Intonation.daughter1.name")))
916
+ (item.feat syl "n.R:Intonation.daughter1.name"))
917
+ ((eq 0 (item.feat syl 'syl_out)) 0) ;;GM real ip_final would be better
918
+ (t (next_accent_type (item.relation.next syl 'Syllable)))))
919
+
920
+ (define (last_accent_type syl)
921
+ "(last_accent_type syl)
922
+ Returns the type of the last (previous) accent."
923
+ (if (>= printdebug 4)
924
+ (begin (print "Entering last_accent_type")))
925
+ (cond
926
+ ((not (equal? "NONE" (item.feat syl 'p.tobi_endtone)))
927
+ (item.feat syl 'R:Syllable.p.tobi_endtone))
928
+ ((not (equal? "NONE" (item.feat syl 'p.tobi_accent)))
929
+ (item.feat syl 'R:Syllable.p.tobi_accent))
930
+ ((eq 0 (item.feat syl 'syl_in)) 0) ;;GM real ip_initial would be better
931
+ (t (last_accent_type (item.prev syl 'Syllable)))))
932
+
933
+ (define (next_accent_start syl)
934
+ "(next_accent_start syl)
935
+ Returns the start time of the vowel of next accented syllable"
936
+ (if (>= printdebug 4)
937
+ (begin (print "Entering next_accent_start")))
938
+ (cond
939
+ ((not (eq 0 (item.feat syl "n.R:Intonation.daughter1.name")))
940
+ (item.feat syl "R:Syllable.n.syllable_start")) ;;GM vowel start would be better
941
+ ((eq 0 (item.feat syl 'syl_out)) 0)
942
+ (t (next_accent_start (item.relation.next syl 'Syllable)))))
943
+
944
+ ; new features (not used yet)
945
+
946
+ (define (ip_final syl)
947
+ "(ip_final SYL)
948
+ returns 1 if the syllable is the final syllable of an
949
+ ip (intermediate phrase)"
950
+ (cond
951
+ ((or (equal? 0 (item.feat syl "syl_out"))
952
+ (equal? "L-" (item.feat syl "tobi_endtone"))
953
+ (equal? "H-" (item.feat syl "tobi_endtone"))
954
+ (equal? "!H-" (item.feat syl "tobi_endtone"))) 1)
955
+ (t 0)))
956
+
957
+ (define (ip_initial syl)
958
+ "(ip_initial SYL)
959
+ returns 1 if the syllable is the initial syllable of an
960
+ ip (intermediate phrase)"
961
+ (cond
962
+ ((equal? 0 (item.feat syl "syl_in"))
963
+ 1)
964
+ ((equal? 1 (ip_final (item.relation.prev syl 'Syllable)))
965
+ 1)
966
+ (t 0)))
967
+
968
+ ;; NEXT TWO FUNCTIONS ARE NEW - MDS
969
+ (define (ip_one_syllable_case syl)
970
+ "(ip_one_syllable_case SYL)
971
+ returns true if the syllable is the initial syllable of an
972
+ ip (intermediate phrase) and doesn't itself contain a complex
973
+ tone that starts opposite this syllable's accent"
974
+ (if (eqv? 0 (ip_initial syl))
975
+ 0
976
+ (let ((accent (item.feat syl "tobi_accent"))
977
+ (tone (item.feat syl "tobi_endtone")))
978
+ (cond
979
+ ((and (equal? tone "L-H%")
980
+ (or (equal? accent "H*")
981
+ (equal? accent "!H*")
982
+ (equal? accent "L+H*")
983
+ (equal? accent "L+!H*")
984
+ (equal? accent "L*+H")
985
+ (equal? accent "L*+!H*")
986
+ (equal? accent "H+!H*")))
987
+ 0)
988
+ ((and (or (equal? tone "H-L%")
989
+ (equal? tone "!H-L%"))
990
+ (equal? accent "L*"))
991
+ 0)
992
+ (t
993
+ 1)))))
994
+
995
+ (define (hstar_weak_target syl)
996
+ (if (and (equal? 0 (item.feat syl 'asyl_in))
997
+ (member (next_accent_type syl)
998
+ (list "L*" "L*+H" "L*+!H" "L+H*" "L+!H*")))
999
+ 1
1000
+ 0))
1001
+
1002
+ (provide 'tobi_rules)
CosyVoice-ttsfrd/resource/festival/token.scm ADDED
@@ -0,0 +1,809 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Various tokenizing functions and customization
35
+
36
+ (define (Token utt)
37
+ "(Token UTT)
38
+ Build a Word stream from the Token stream, analyzing compound words
39
+ numbers etc as tokens into words. Respects the Parameter Language
40
+ to choose the appropriate token to word module."
41
+ (let ((rval (apply_method 'Token_Method utt)) ;; might be defined
42
+ (language (Parameter.get 'Language)))
43
+ (cond
44
+ (rval rval) ;; newer style
45
+ ((or (string-equal "britishenglish" language)
46
+ (string-equal "english" language)
47
+ (string-equal "americanenglish" language))
48
+ (Token_English utt))
49
+ ((string-equal "welsh" language)
50
+ (Token_Welsh utt))
51
+ (t
52
+ (Token_Any utt)))))
53
+
54
+ (define (remove_leadtrail_underscores name)
55
+ "(remove_leadtrail_underscores name)
56
+ Get rid of leading and trailing underscores that may be used for emphasis,
57
+ not this is called when there are underscores at the beginning and end but
58
+ there may not be an equal number of them."
59
+ (let ((se (symbolexplode name)))
60
+ (while (string-equal "_" (car se))
61
+ (set! se (cdr se)))
62
+ (set! se (reverse se))
63
+ (while (string-equal "_" (car se))
64
+ (set! se (cdr se)))
65
+ (apply string-append (reverse se))))
66
+
67
+ (define (english_token_to_words token name)
68
+ "(english_token_to_words TOKEN NAME)
69
+ Returns a list of words for NAME from TOKEN. This allows the
70
+ user to customize various non-local, multi-word, context dependent
71
+ translations of tokens into words. If this function is unset only
72
+ the builtin translation rules are used, if this is set the builtin
73
+ rules are not used unless explicitly called. [see Token to word rules]"
74
+ (cond
75
+ ((string-equal name "POA")
76
+ (list "P" "O" (list '(name "A") (list 'pos token.letter_pos))))
77
+ ((string-equal name "US")
78
+ (list "U" "S"))
79
+ ((string-equal name "AI")
80
+ (list (list '(name "A") (list 'pos token.letter_pos)) "I"))
81
+ ((string-equal (downcase name) "openai")
82
+ (list "Open" (list '(name "A") (list 'pos token.letter_pos)) "I"))
83
+ ((string-equal name "EB")
84
+ (list "E" "B"))
85
+ ((string-equal (downcase name) "ios")
86
+ (list "I" "O" "S"))
87
+ ((string-equal (downcase name) "content")
88
+ (if (and (string-equal (downcase (item.feat token "n.name")) "with")
89
+ (string-matches (downcase (item.feat token "p.name")) "be\\|been\\|being\\|is\\(n'?t\\)?\\|are\\(n'?t\\)?\\|was\\(n'?t\\)\\|were\\(n'?t\\)\\|not\\|will\\|won'?t\\|may\\|would\\(n'?t\\)?\\|can\\(not\\)?\\|could\\(n'?t\\)?\\|should\\(n'?t\\)?\\|must\\(n'?t\\)?\\|might"))
90
+ (list (list (list 'name name)(list 'pos 'jj)))
91
+ (list (list (list 'name name)(list 'pos 'nn)))))
92
+ ((string-equal name "/") (list ""))
93
+ ((and (string-equal (downcase name) "refuse")
94
+ (string-equal (downcase (item.feat token "n.name")) "to"))
95
+ (list (list (list 'name name)(list 'pos 'vb))))
96
+ ((string-equal (downcase name) "live")
97
+ (cond
98
+ ((or (string-matches (downcase (item.feat token "n.name")) "nation\\|tv\\|news\\|broadcasts?\\|streams?\\|stream\\(ing\\)?\\|shows?\\|events?\\|games?\\|match\\(es\\)?\\|scores?")
99
+ (string-matches (downcase (item.feat token "nn.name")) "games?\\|match\\(es\\)?\\|scores?"))
100
+ (list (list (list 'name name)(list 'pos 'jj))))
101
+ ((string-equal (item.feat token "p.name") "I")
102
+ (list (list (list 'name name)(list 'pos 'vb))))
103
+ ((not (string-equal (item.feat token "punc") 0))
104
+ (if (string-matches (downcase (item.feat token "p.name")) "to\\|I\\|we\\|you\\|they\\|he\\|she\\|will\\|may\\|would\\|can\\|could\\|should\\|must\\|might")
105
+ (list (list (list 'name name)(list 'pos 'vb)))
106
+ (list (list (list 'name name)(list 'pos 'jj)))))
107
+ (t
108
+ (builtin_english_token_to_words token name))))
109
+ ((string-equal name "#") (list "pound"))
110
+ ((string-equal name "$") (list "dollar"))
111
+ ((string-matches name "[A-Z]*[\\$#][0-9,]+\\(\\.[0-9]+\\)?")
112
+ ;; Some form of money (pounds or type of dollars)
113
+ (let (amount type currency)
114
+ (cond
115
+ ((string-matches name ".*\\$.*")
116
+ (set! amount (string-after name "$"))
117
+ (set! type (string-before name "$"))
118
+ (set! currency "dollar"))
119
+ ((string-matches name ".*#.*")
120
+ (set! amount (string-after name "#"))
121
+ (set! type (string-before name "#"))
122
+ (set! currency "pound"))
123
+ (t
124
+ ;; who knows
125
+ (set! amount (string-after name "$"))
126
+ (set! type (string-before name "$"))
127
+ (set! currency "dollar")))
128
+ (cond
129
+ ((string-matches (item.feat token "n.name")
130
+ ".*illion.?")
131
+ (append ;; "billions and billions" - Sagan
132
+ (builtin_english_token_to_words token amount)
133
+ (list (item.feat token "n.name")) ;; illion
134
+ (token_money_expand type)
135
+ (list (string-append currency "s"))))
136
+ ((string-matches amount ".*\\...$")
137
+ (cond
138
+ ((string-equal amount "0.00")
139
+ (append
140
+ (list "zero")
141
+ (list (string-append currency "s"))))
142
+ ((string-equal amount "0.01")
143
+ (append
144
+ (list "one")
145
+ (if (string-equal currency "dollar")
146
+ (list "cent")
147
+ (list "penny"))))
148
+ ((string-matches amount "0\\.[0-9][0-9]")
149
+ (append
150
+ (builtin_english_token_to_words
151
+ token (remove_leading_zeros (string-after amount ".")))
152
+ (if (string-equal currency "dollar")
153
+ (list "cents")
154
+ (list "pence"))))
155
+ (t
156
+ (append
157
+ (builtin_english_token_to_words token (string-before amount "."))
158
+ (token_money_expand type)
159
+ (if (string-matches amount "1\\..*")
160
+ (list currency)
161
+ (list (string-append currency "s")))
162
+ (cond
163
+ ((string-matches name ".*\\.00$")
164
+ nil)
165
+ ((string-matches name ".*\\.01$")
166
+ (append
167
+ (list "and" "one")
168
+ (if (string-equal currency "dollar")
169
+ (list "cent")
170
+ (list "penny"))))
171
+ (t
172
+ (append
173
+ (list "and")
174
+ (builtin_english_token_to_words
175
+ token (remove_leading_zeros (string-after amount ".")))
176
+ (if (string-equal currency "dollar")
177
+ (list "cents")
178
+ (list "pence")))))))))
179
+ (t
180
+ (append ;; nothing after point or lots after point
181
+ (builtin_english_token_to_words token amount)
182
+ (token_money_expand type)
183
+ (if (or (string-matches amount "1")
184
+ (string-equal currency "yuan"))
185
+ (list currency)
186
+ (list (string-append currency "s"))))))))
187
+ ((and (string-matches name ".*illion.?")
188
+ (string-matches (item.feat token "p.name")
189
+ "[A-Z]*[\\$#][0-9,]+\\(\\.[0-9]+\\)?"))
190
+ nil ;; dealt with on the previous symbol
191
+ )
192
+ ((string-matches name "[1-9][0-9]*/[1-9][0-9]*")
193
+ (let ((numerator (string-before name "/"))
194
+ (denominator (string-after name "/"))
195
+ )
196
+ (cond
197
+ ((string-matches name "1/2")
198
+ (list "half"))
199
+ ((string-matches denominator "4")
200
+ (append
201
+ (builtin_english_token_to_words token numerator)
202
+ (list "quarter")
203
+ (if (string-equal numerator "1")
204
+ (list '((name "'s")(pos nnp)))
205
+ nil)))
206
+ (t
207
+ (append
208
+ (builtin_english_token_to_words token numerator)
209
+ (begin
210
+ (item.set_feat token "token_pos" "ordinal")
211
+ (builtin_english_token_to_words token denominator))
212
+ (if (string-equal numerator "1")
213
+ nil
214
+ (list '((name "'s")(pos nnp)))))))))
215
+ ((and (string-matches name "No")
216
+ (item.next token)
217
+ (string-matches (item.feat token "n.name")
218
+ "[0-9]+"))
219
+ (list
220
+ "number"))
221
+ ((string-matches name ".*%$")
222
+ (append
223
+ (token_to_words token (string-before name "%"))
224
+ (list "percent")))
225
+ ((string-matches name "[0-9]+s") ;; e.g. 1950s
226
+ (item.set_feat token "token_pos" "year") ;; reasonable guess
227
+ (append
228
+ (builtin_english_token_to_words token (string-before name "s"))
229
+ (list '((name "'s")(pos nnp))) ;; will get assimilated by postlexical rules
230
+ ))
231
+ ((string-matches name "[0-9]+'s") ;; e.g. 1950's
232
+ (item.set_feat token "token_pos" "year") ;; reasonable guess
233
+ (append
234
+ (builtin_english_token_to_words token (string-before name "'s"))
235
+ (list '((name "'s")(pos nnp))) ;; will get assimilated by postlexical rules
236
+ ))
237
+ ((and (string-matches name ".*s$")
238
+ (string-equal (item.feat token "punc") "'"))
239
+ ;; potential possessive or may be end of a quote
240
+ (if (token_no_starting_quote token)
241
+ (item.set_feat token "punc" ""))
242
+ (builtin_english_token_to_words token name))
243
+ ((and (string-equal name "A") ;; letter or determiner
244
+ (or (and (string-matches (item.feat token "p.name") "[A-Za-z].*")
245
+ (or (string-equal (item.feat token "p.punc") 0)
246
+ (not string-matches (item.feat token "p.punc") ".*[?!:\\.;].*")))
247
+ (string-matches (item.feat token "n.name") "[A-Z].*")))
248
+ (list (list '(name "a")(list 'pos token.letter_pos))))
249
+ ((and (or (string-equal name "A") (string-equal name "a")) ;; letter or determiner
250
+ (and (string-equal (item.feat token "p.name") 0)
251
+ (string-equal (item.feat token "n.name") 0)))
252
+ (list (list '(name "a")(list 'pos token.letter_pos))))
253
+ ((member_string name english_homographs)
254
+ (list (list (list 'name name)
255
+ (list 'hg_pos (item.feat token "token_pos")))))
256
+ ((string-matches name "__*[^_][^_]*_*_") ;; _emphasis_
257
+ (english_token_to_words
258
+ token
259
+ (remove_leadtrail_underscores name)
260
+ ))
261
+ ((string-matches name "\\(EW\\|NS\\|NE\\|CC\\|DT\\|TE\\|JS\\|JW\\|JE\\|CR\\|CP\\)[1-4][0-9]") ;; Singapore MRT station IDs
262
+ (let (lname num)
263
+ (cond
264
+ ((string-matches name "EW.*")
265
+ (set! lname "EW"))
266
+ ((string-matches name "NS.*")
267
+ (set! lname "NS"))
268
+ ((string-matches name "NE.*")
269
+ (set! lname "NE"))
270
+ ((string-matches name "CC.*")
271
+ (set! lname "CC"))
272
+ ((string-matches name "DT.*")
273
+ (set! lname "DT"))
274
+ ((string-matches name "TE.*")
275
+ (set! lname "TE"))
276
+ ((string-matches name "JS.*")
277
+ (set! lname "JS"))
278
+ ((string-matches name "JW.*")
279
+ (set! lname "JW"))
280
+ ((string-matches name "JE.*")
281
+ (set! lname "JE"))
282
+ ((string-matches name "CR.*")
283
+ (set! lname "CR"))
284
+ (t
285
+ (set! lname "CP")))
286
+ (set! num (string-after name lname))
287
+ (append
288
+ (symbolexplode lname)
289
+ (builtin_english_token_to_words token num))))
290
+ ((string-matches name "\\([1-9]\\|10\\|11\\|12\\)[AaPp][Mm]") ;; time
291
+ (let (hour apm (atime (downcase name)))
292
+ (if (string-matches atime ".*am")
293
+ (set! apm "am")
294
+ (set! apm "pm"))
295
+ (set! hour (string-before atime apm))
296
+ (append
297
+ (builtin_english_token_to_words token hour)
298
+ (if (string-equal apm "am")
299
+ (builtin_english_token_to_words token "A.M")
300
+ (builtin_english_token_to_words token "P.M")))))
301
+ ((string-matches name "[0-9]?[0-9][:\\.][0-9][0-9][AaPp][Mm]") ;; time
302
+ ;; must be am/pm present for . to be acceptable separator
303
+ (let (hours mins half sep (ttime (downcase name)))
304
+ (if (string-matches ttime ".*:.*")
305
+ (set! sep ":")
306
+ (set! sep "."))
307
+ (set! hours (string-before ttime sep))
308
+ (set! mins (string-after ttime sep))
309
+ (if (string-matches ttime ".*am")
310
+ (set! sep "am")
311
+ (set! sep "pm"))
312
+ (set! mins (string-before mins sep))
313
+ (append
314
+ (builtin_english_token_to_words token hours)
315
+ (cond
316
+ ((string-equal mins "00")
317
+ nil)
318
+ ((string-matches mins "0.")
319
+ (cons
320
+ "oh"
321
+ (builtin_english_token_to_words token (string-after mins "0"))))
322
+ (t
323
+ (builtin_english_token_to_words token mins)))
324
+ (if (string-equal sep "am")
325
+ (builtin_english_token_to_words token "A.M")
326
+ (builtin_english_token_to_words token "P.M")))))
327
+ ((string-matches name "[0-9]?[0-9]:[0-9][0-9]") ;; time
328
+ (append
329
+ (builtin_english_token_to_words
330
+ token (remove_leading_zeros (string-before name ":")))
331
+ (cond
332
+ ((string-equal "00" (string-after name ":"))
333
+ nil)
334
+ ((string-matches (string-after name ":") "0.")
335
+ (cons
336
+ "oh"
337
+ (builtin_english_token_to_words
338
+ token
339
+ (remove_leading_zeros (string-after name ":")))))
340
+ (t
341
+ (builtin_english_token_to_words
342
+ token
343
+ (string-after name ":"))))))
344
+ ((string-matches name "[0-9][0-9]:[0-9][0-9]:[0-9][0-9]") ;; exact time
345
+ (append
346
+ (builtin_english_token_to_words
347
+ token (remove_leading_zeros (string-before name ":")))
348
+ (list "hours")
349
+ (builtin_english_token_to_words
350
+ token (remove_leading_zeros
351
+ (string-before (string-after name ":") ":")))
352
+ (list "minutes" "and")
353
+ (builtin_english_token_to_words
354
+ token (remove_leading_zeros
355
+ (string-after (string-after name ":") ":")))
356
+ (list "seconds")))
357
+ ((string-matches name "[12][0-9][0-9][0-9]/\\(1[012]\\|0[1-9]\\)/\\(0[1-9]\\|[12][0-9]\\|3[01]\\)")
358
+ (let ((year (string-before name "/"))
359
+ (num1 (string-before (string-after name "/") "/"))
360
+ (num2 (string-after (string-after name "/") "/"))
361
+ month day)
362
+ (cond
363
+ ((string-equal num1 "01")
364
+ (set! month "january"))
365
+ ((string-equal num1 "02")
366
+ (set! month "february"))
367
+ ((string-equal num1 "03")
368
+ (set! month "march"))
369
+ ((string-equal num1 "04")
370
+ (set! month "april"))
371
+ ((string-equal num1 "05")
372
+ (set! month "may"))
373
+ ((string-equal num1 "06")
374
+ (set! month "june"))
375
+ ((string-equal num1 "07")
376
+ (set! month "july"))
377
+ ((string-equal num1 "08")
378
+ (set! month "august"))
379
+ ((string-equal num1 "09")
380
+ (set! month "september"))
381
+ ((string-equal num1 "10")
382
+ (set! month "october"))
383
+ ((string-equal num1 "11")
384
+ (set! month "november"))
385
+ ((string-equal num1 "12")
386
+ (set! month "december"))
387
+ (t
388
+ (set! month (tok_string_as_letters num1))))
389
+ (item.set_feat token "token_pos" "ordinal")
390
+ (set! day (builtin_english_token_to_words token num2))
391
+ (item.set_feat token "token_pos" "year")
392
+ (append
393
+ (list month)
394
+ (list "the")
395
+ day
396
+ (builtin_english_token_to_words token year))))
397
+ ((string-matches name "[0-9][0-9]?/[0-9][0-9]?/[0-9][0-9]\\([0-9][0-9]\\)?")
398
+ ;; date, say it as numbers to avoid American/British problem
399
+ (let ((num1 (string-before name "/"))
400
+ (num2 (string-before (string-after name "/") "/"))
401
+ (year (string-after (string-after name "/") "/"))
402
+ day month)
403
+ (item.set_feat token "token_pos" "cardinal")
404
+ (set! day (builtin_english_token_to_words token num1))
405
+ (set! month (builtin_english_token_to_words token num2))
406
+ (item.set_feat token "token_pos" "year")
407
+ (append
408
+ day
409
+ month
410
+ (list '((name ",")(pbreak_scale 0.9)))
411
+ (builtin_english_token_to_words token year))))
412
+ ((string-matches name "[0-9][0-9][0-9]-[0-9][0-9][0-9][0-9]")
413
+ (item.set_feat token "token_pos" "digits") ;; canonical phone number
414
+ (append
415
+ (builtin_english_token_to_words token (string-before name "-"))
416
+ (list '((name ",")(pbreak_scale 1.0)))
417
+ (builtin_english_token_to_words token (string-after name "-"))))
418
+ ((string-matches name "[0-9]+-[0-9]+-[-0-9]+")
419
+ ;; long distance number
420
+ (let ((r '(dummy)) (remainder name))
421
+ (item.set_feat token "token_pos" "digits")
422
+ (while (> (length remainder) 0)
423
+ (if (string-matches remainder "[0-9]+")
424
+ (set! r (append r
425
+ (builtin_english_token_to_words
426
+ token remainder)))
427
+ (set! r (append r
428
+ (builtin_english_token_to_words
429
+ token (string-before remainder "-")))))
430
+ (set! remainder (string-after remainder "-"))
431
+ (if (> (length remainder) 0)
432
+ (set! r (append r (list '((name ",")(pbreak_scale 1.0)))))))
433
+ (cdr r))
434
+ )
435
+ ((and (string-matches name "[0-9][0-9][0-9]")
436
+ (string-matches (item.feat token "n.name")
437
+ "[0-9][0-9][0-9]-[0-9][0-9][0-9][0-9]"))
438
+ (item.set_feat token "token_pos" "digits")
439
+ (builtin_english_token_to_words token name))
440
+ ((string-matches name "[0-9]+-[0-9]+")
441
+ (let ((tokpos))
442
+ (item.set_name token (string-before name "-"))
443
+ (set! tokpos (wagon token
444
+ (car (cdr (assoc "[0-9]+" token_pos_cart_trees)))))
445
+ (item.set_feat token "token_pos" (car tokpos))
446
+ (append
447
+ (builtin_english_token_to_words token (string-before name "-"))
448
+ (list "to")
449
+ (builtin_english_token_to_words token (string-after name "-")))))
450
+ ((string-matches name "\\(iii?\\|II?I?\\|IV\\|VI?I?I?\\|IX\\|X[VIX]*\\)")
451
+ ;; Roman numerals
452
+ (let ((tp (item.feat token "token_pos")))
453
+ (cond
454
+ ((string-matches tp "century");; always believe this
455
+ (item.set_feat token "token_pos" "ordinal")
456
+ (if (or (string-equal "1" (tok_rex token))
457
+ (item.feat token "p.lisp_tok_rex_names"))
458
+ (append
459
+ (list "the")
460
+ (builtin_english_token_to_words
461
+ token (tok_roman_to_numstring name)))
462
+ (builtin_english_token_to_words
463
+ token (tok_roman_to_numstring name))))
464
+ ((string-matches name "[IVX]");; be *very* wary of this one
465
+ (if (and (string-equal
466
+ "1" (item.feat token "p.lisp_tok_section_name"))
467
+ (string-matches tp "number"))
468
+ (builtin_english_token_to_words
469
+ token (tok_roman_to_numstring name))
470
+ (tok_string_as_letters name)))
471
+ ((or (string-matches tp "number")
472
+ (string-matches name "iii?\\|III?"))
473
+ (item.set_feat token "token_pos" "cardinal")
474
+ (builtin_english_token_to_words
475
+ token (tok_roman_to_numstring name)))
476
+ (t;; else its a letter
477
+ (tok_string_as_letters name)))))
478
+ ((and (string-matches name "pp")
479
+ (string-matches (item.feat token "n.name")
480
+ "[0-9]+-[0-9]+"))
481
+ (list "pages"))
482
+ ((and (string-matches name "ss")
483
+ (string-matches (item.feat token "n.name")
484
+ "[0-9]+-[0-9]+"))
485
+ (list "sections"))
486
+ ((string-matches name "_____+")
487
+ (list "line" "of" "underscores"))
488
+ ((string-matches name "=====+")
489
+ (list "line" "of" "equals"))
490
+ ((string-matches name "-----+")
491
+ (list "line" "of" "hyphens"))
492
+ ((string-matches name "\\*\\*\\*\\*\\*+")
493
+ (list "line" "of" "asterisks"))
494
+ ((string-matches name "--+")
495
+ (list '((name ",")(pbreak_scale 1.0))))
496
+ ((string-matches name ".*--+.*")
497
+ (append
498
+ (builtin_english_token_to_words token (string-before name "--"))
499
+ (list '((name ",")(pbreak_scale 1.0)))
500
+ (builtin_english_token_to_words token (string-after name "--"))))
501
+ ((string-matches name "[A-Z][A-Z]?&[A-Z][A-Z]?")
502
+ (append
503
+ (tok_string_as_letters (string-before name "&"))
504
+ (list "and")
505
+ (tok_string_as_letters (string-after name "&"))))
506
+ ((and (string-equal name "Ms")
507
+ (string-matches (item.feat token "n.name") "[A-Z][^A-Z]*"))
508
+ (list "mizz"))
509
+ ((string-matches name "[A-Z]+'s")
510
+ (append
511
+ (builtin_english_token_to_words token (string-before name "'s"))
512
+ (list '((name "'s")(pos nnp))) ;; will get assimilated by postlexical rules
513
+ ))
514
+ ((or (string-matches name "[A-Z][A-Z]+s")
515
+ (string-matches name "[BCDEFGHJKLMNOPQRSTVWXYZ]+s"))
516
+ (append
517
+ (builtin_english_token_to_words token (string-before name "s"))
518
+ (list '((name "'s")(pos nnp))) ;; will get assimilated by postlexical rules
519
+ ))
520
+ ((string-matches name "<.*@.*>") ;; quoted e-mail
521
+ (append
522
+ (builtin_english_token_to_words
523
+ token (string-after (string-before name "@") "<"))
524
+ (list "at")
525
+ (builtin_english_token_to_words
526
+ token (string-before (string-after name "@") ">"))))
527
+ ((string-matches name "\\(Dr\\|St\\)")
528
+ (if (string-equal (item.feat token "token_pos") "street")
529
+ (if (string-matches name "Dr")
530
+ (list "drive")
531
+ (list "street"))
532
+ (if (string-matches name "Dr") ;; default on title side
533
+ (list "doctor")
534
+ (list "saint"))))
535
+ ((string-matches name "[Cc]alif") ;; hopelessly specific ...
536
+ (list
537
+ "california"))
538
+ (t
539
+ (builtin_english_token_to_words token name))))
540
+
541
+ ;;; This is set as the default
542
+ (defvar token_to_words english_token_to_words)
543
+
544
+ (defvar token.punctuation "\"'`.,:;!?(){}[]"
545
+ "token.punctuation
546
+ A string of characters which are to be treated as punctuation when
547
+ tokenizing text. Punctuation symbols will be removed from the text
548
+ of the token and made available through the \"punctuation\" feature.
549
+ [see Tokenizing]")
550
+ (defvar token.prepunctuation "\"'`({["
551
+ "token.prepunctuation
552
+ A string of characters which are to be treated as preceding punctuation
553
+ when tokenizing text. Prepunctuation symbols will be removed from the text
554
+ of the token and made available through the \"prepunctuation\" feature.
555
+ [see Tokenizing]")
556
+ (defvar token.whitespace " \t\n\r "
557
+ "token.whitespace
558
+ A string of characters which are to be treated as whitespace when
559
+ tokenizing text. Whitespace is treated as a separator and removed
560
+ from the text of a token and made available through the \"whitespace\"
561
+ feature. [see Tokenizing]")
562
+ (defvar token.singlecharsymbols ""
563
+ "token.singlecharsymbols
564
+ Characters which have always to be split as tokens. This would be
565
+ usual is standard text, but is useful in parsing some types of
566
+ file. [see Tokenizing]")
567
+
568
+ (defvar token.letter_pos 'nn
569
+ "token.letter_pos
570
+ The part of speech tag (valid for your part of speech tagger) for
571
+ individual letters. When the tokenizer decide to pronounce a token
572
+ as a list of letters this tag is added to each letter in the list.
573
+ Note this should be from the part of speech set used in your tagger
574
+ which may not be the same one that appears in the actual lexical
575
+ entry (if you map them afterwards). This specifically allows \"a\"
576
+ to come out as ae rather than @.")
577
+
578
+ (defvar token.unknown_word_name "unknown"
579
+ "token.unknown_word_name
580
+ When all else fails and a pronunciation for a word or character can't
581
+ be found this word will be said instead. If you make this \"\" them
582
+ the unknown word will simple be omitted. This will only
583
+ really be called when there is a bug in the lexicon and characters
584
+ are missing from the lexicon. Note this word should be in the lexicon.")
585
+
586
+ (def_feature_docstring
587
+ 'Token.punc
588
+ "Token.punc
589
+ Succeeding punctuation symbol found after token in original
590
+ string/file.")
591
+ (def_feature_docstring
592
+ 'Token.whitespace
593
+ "Token.whitespace
594
+ Whitespace found before token in original string/file.")
595
+ (def_feature_docstring
596
+ 'Token.prepunctuation
597
+ "Token.prepunctuation
598
+ Preceeding puctuation symbol found before token in original string/file.")
599
+
600
+ (require 'tokenpos)
601
+ ;;;
602
+ ;;; Token pos are gross level part of speech tags which help decide
603
+ ;;; pronunciation of tokens (particular expansion of Tokens into words)
604
+ ;;; The most obvious example is identifying number types (ordinals,
605
+ ;;; years, digits or numbers).
606
+ ;;;
607
+ (defvar english_token_pos_cart_trees
608
+ '(
609
+ ;; Format is (Regex Tree)
610
+ ("[0-9]+"
611
+ ((lisp_num_digits < 3.8)
612
+ ((p.lisp_token_pos_guess is month)
613
+ ((lisp_month_range is 0) ((cardinal)) ((ordinal)))
614
+ ((n.lisp_token_pos_guess is month)
615
+ ((lisp_month_range is 0) ((cardinal)) ((ordinal)))
616
+ ((n.lisp_token_pos_guess is numeric)
617
+ ((lisp_num_digits < 2)
618
+ ((p.lisp_token_pos_guess is numeric)
619
+ ((pp.lisp_token_pos_guess is sym) ((digits)) ((cardinal)))
620
+ ((cardinal)))
621
+ ((nn.lisp_token_pos_guess is sym) ((cardinal)) ((digits))))
622
+ ((lisp_num_digits < 2)
623
+ ((nn.lisp_token_pos_guess is numeric)
624
+ ((n.lisp_token_pos_guess is sym)
625
+ ((lisp_month_range is 0) ((digits)) ((cardinal)))
626
+ ((cardinal)))
627
+ ((cardinal)))
628
+ ((name < 302.3)
629
+ ((p.lisp_token_pos_guess is flight)
630
+ ((digits))
631
+ ((n.lisp_token_pos_guess is sym)
632
+ ((p.lisp_token_pos_guess is sym) ((digits)) ((cardinal)))
633
+ ((cardinal))))
634
+ ((p.lisp_token_pos_guess is a)
635
+ ((digits))
636
+ ((n.lisp_token_pos_guess is sym)
637
+ ((nn.lisp_token_pos_guess is sym)
638
+ ((name < 669.2) ((digits)) ((cardinal)))
639
+ ((cardinal)))
640
+ ((name < 373.2)
641
+ ((cardinal))
642
+ ((name < 436.2)
643
+ ((name < 392.6) ((digits)) ((cardinal)))
644
+ ((name < 716.5)
645
+ ((cardinal))
646
+ ((name < 773.6)
647
+ ((p.lisp_token_pos_guess is _other_) ((digits)) ((cardinal)))
648
+ ((cardinal)))))))))))))
649
+ ((p.lisp_token_pos_guess is numeric)
650
+ ((pp.lisp_token_pos_guess is month)
651
+ ((year))
652
+ ((nn.lisp_token_pos_guess is numeric)
653
+ ((cardinal))
654
+ ((p.lisp_token_pos_precise_val is year)
655
+ ((cardinal))
656
+ ((p.lisp_token_pos_precise_val is ordinal)
657
+ ((cardinal))
658
+ ((digits))))))
659
+ ((nn.lisp_token_pos_guess is numeric)
660
+ ((n.lisp_token_pos_guess is month)
661
+ ((cardinal))
662
+ ((n.lisp_token_pos_guess is numeric)
663
+ ((digits))
664
+ ((p.lisp_token_pos_guess is _other_) ((cardinal)) ((year)))))
665
+ ((p.lisp_token_pos_guess is _other_)
666
+ ((lisp_num_digits < 4.4)
667
+ ((name < 2959.6)
668
+ ((name < 1773.4) ((cardinal)) ((year)))
669
+ ((cardinal)))
670
+ ((pp.lisp_token_pos_guess is _other_) ((digits)) ((cardinal))))
671
+ ((n.lisp_token_pos_guess is to)
672
+ ((year))
673
+ ((p.lisp_token_pos_guess is sym)
674
+ ((pp.lisp_token_pos_guess is sym)
675
+ ((cardinal))
676
+ ((lisp_num_digits < 4.6) ((year)) ((digits))))
677
+ ((lisp_num_digits < 4.8)
678
+ ((name < 2880)
679
+ ((name < 1633.2)
680
+ ((name < 1306.4) ((cardinal)) ((year)))
681
+ ((year)))
682
+ ((cardinal)))
683
+ ((cardinal)))))))))
684
+ )
685
+ ("\\(II?I?\\|IV\\|VI?I?I?\\|IX\\|X[VIX]*\\)";; Roman numerals
686
+ ((p.lisp_tok_rex_names is 0)
687
+ ((lisp_num_digits is 5)
688
+ ((number))
689
+ ((lisp_num_digits is 4)
690
+ ((number))
691
+ ((nn.lisp_num_digits is 13)
692
+ ((number))
693
+ ((p.lisp_num_digits is 7)
694
+ ((number))
695
+ ((p.lisp_tok_section_name is 0)
696
+ ((lisp_tok_rex is 0)
697
+ ((lisp_num_digits is 3)
698
+ ((p.lisp_num_digits is 4)
699
+ ((number))
700
+ ((nn.lisp_num_digits is 4)
701
+ ((number))
702
+ ((n.lisp_num_digits is 4)
703
+ ((number))
704
+ ((pp.lisp_num_digits is 3)
705
+ ((number))
706
+ ((p.lisp_num_digits is 2)
707
+ ((letter))
708
+ ((nn.lisp_num_digits is 2)
709
+ ((letter))
710
+ ((n.cap is 0) ((letter)) ((number)))))))))
711
+ ((nn.lisp_num_digits is 11)
712
+ ((letter))
713
+ ((lisp_num_digits is 1)
714
+ ((pp.lisp_num_digits is 9)
715
+ ((letter))
716
+ ((p.lisp_num_digits is 9)
717
+ ((letter))
718
+ ((n.lisp_num_digits is 6)
719
+ ((letter))
720
+ ((pp.lisp_num_digits is 6)
721
+ ((letter))
722
+ ((pp.cap is 0)
723
+ ((n.cap is 0)
724
+ ((p.lisp_num_digits is 1)
725
+ ((letter))
726
+ ((n.lisp_num_digits is 4) ((letter)) ((letter))))
727
+ ((letter)))
728
+ ((letter)))))))
729
+ ((p.lisp_num_digits is 10)
730
+ ((number))
731
+ ((n.lisp_num_digits is 8)
732
+ ((number))
733
+ ((pp.lisp_num_digits is 9)
734
+ ((number))
735
+ ((nn.lisp_num_digits is 5)
736
+ ((number))
737
+ ((n.lisp_num_digits is 4) ((number)) ((letter))))))))))
738
+ ((letter)))
739
+ ((number)))))))
740
+ ((century))))
741
+ ("\\(Dr\\|St\\)"
742
+ ((n.name is 0)
743
+ ((p.cap is 1)
744
+ ((street))
745
+ ((p.name matches "[0-9]*\\(1[sS][tT]\\|2[nN][dD]\\|3[rR][dD]\\|[0-9][tT][hH]\\)")
746
+ ((street))
747
+ ((title))))
748
+ ((punc matches ".*,.*")
749
+ ((street))
750
+ ((p.punc matches ".*,.*")
751
+ ((title))
752
+ ((n.cap is 0)
753
+ ((street))
754
+ ((p.cap is 0)
755
+ ((p.name matches "[0-9]*\\(1[sS][tT]\\|2[nN][dD]\\|3[rR][dD]\\|[0-9][tT][hH]\\)")
756
+ ((street))
757
+ ((title)))
758
+ ((pp.name matches "[1-9][0-9]+")
759
+ ((street))
760
+ ((title)))))))))
761
+ ("lead"
762
+ ((p.name in (was were have had been having has is are))
763
+ ((led))
764
+ ((liid))))
765
+ ("read"
766
+ ((p.name in ("was" "wasn't" "were" "weren't" "have" "haven't" "had" "hadn't" "been" "having" "has" "hasn't" "is" "isn't" "are" "aren't" "He" "She"))
767
+ ((red))
768
+ ((riid))))
769
+ ;("read"
770
+ ; ((p.name matches "[Pp]lease\\|PLEASE\\|[Ll]et'?s\\|LET'?S\\|[Tt][Oo]\\|[Ww]ill\\|WILL\\|[Ww]on'?t\\|WON'?T\\|[Mm]ay\\|MAY\\|[Cc]an\\(not\\|'?t\\)?\\|CAN\\(NOT\\|'?T\\)?\\|[Mm]ight\\|MIGHT\\|\\([Ww]ould\\|[Cc]ould\\|[Ss]hould\\|[Mm]ust\\)\\(n'?t\\)?\\|\\(WOULD\\|COULD\\|SHOULD\\|MUST\\)\\(N'?T\\)?")
771
+ ; ((riid))
772
+ ; ((p.name is 0)
773
+ ; ((riid))
774
+ ; ((red)))))
775
+ ("and"
776
+ ((n.name is 0)
777
+ ((aend))
778
+ ((punc is 0)
779
+ ((ahnd))
780
+ ((aend)))))
781
+ ("for"
782
+ ((n.name is 0)
783
+ ((faor))
784
+ ((punc is 0)
785
+ ((fer))
786
+ ((faor)))))
787
+ ("to"
788
+ ((n.name is 0)
789
+ ((tuu))
790
+ ((punc is 0)
791
+ ((to))
792
+ ((tuu)))))
793
+ ))
794
+
795
+ (defvar english_homographs
796
+ '("lead" "read" "and" "for" "to")
797
+ "english_homographs
798
+ A list of tokens that are dealt with by a homograph disambiguation tree
799
+ in english_token_pos_cart_trees.")
800
+
801
+ (defvar token_pos_cart_trees
802
+ english_token_pos_cart_trees
803
+ "token_pos_cart_trees
804
+ This is a list of pairs or regex plus CART tree. Tokens that match
805
+ the regex will have the CART tree aplied, setting the result as
806
+ the token_pos feature on the token. The list is checked in order
807
+ and only the first match will be applied.")
808
+
809
+ (provide 'token)
CosyVoice-ttsfrd/resource/festival/tokenpos.scm ADDED
@@ -0,0 +1,286 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Functions used in identifying token types.
35
+ ;;;
36
+
37
+ (defvar token_most_common
38
+ '(
39
+ sym numeric month to day in the of on and writes a years from
40
+ for jst at million by is was gmt page he that than more since as when
41
+ with but after about or his i has it date no died number bst who miles
42
+ university some people an only w year have ago were are pages up days
43
+ months hours minutes through out had which least hi last now ft this
44
+ all one its there between cents until over will before past they
45
+ nearly times tim message so lbs just if age we during she billion then
46
+ other be time new her first states not you members under would many
47
+ says degrees two next fax week while bush been around including back
48
+ campaign american within publisher flight points even early later
49
+ world countries every edt can president most could their what them
50
+ former began women killed another also received long americans pounds
51
+ do dear said km made into did dead war tel still old x took total men
52
+ like f am less c well late down weeks end chapter among place house
53
+ away him election death almost students state soviet where version
54
+ summer man s nation because washington top though m id est these spent
55
+ seats gnu estimated those lost ian high each copies children acres
56
+ tons son per my found won off seconds power nations federal born
57
+ presidential much city begin p name different whose three home hello
58
+ million-euro)
59
+
60
+ "token_most_common
61
+ A list of (English) words which were found to be most common in
62
+ an text database and are used as discriminators in token analysis.")
63
+
64
+ (define (token_pos_precise_val sc)
65
+ "(tok_pos sc)
66
+ Returns a general pos for sc's name.
67
+ numeric All digits
68
+ number float or comma'd numeric
69
+ sym Contains at least one non alphanumeric
70
+ month has month name (or abbrev)
71
+ day has day name (or abbrev)
72
+ rootname else downcased alphabetic.
73
+ Note this can be used to find token_pos but isn't used directly as
74
+ its not disciminatory enough."
75
+ (let ((name (downcase (item.feat sc "token_pos"))))
76
+ (cond
77
+ ((string-equal name "year")
78
+ 'year)
79
+ ((string-equal name "ordinal")
80
+ 'ordinal)
81
+ (t
82
+ '_other_))))
83
+
84
+ (define (token_pos_guess sc)
85
+ "(tok_pos sc)
86
+ Returns a general pos for sc's name.
87
+ numeric All digits
88
+ number float or comma'd numeric
89
+ sym Contains at least one non alphanumeric
90
+ month has month name (or abbrev)
91
+ day has day name (or abbrev)
92
+ rootname else downcased alphabetic.
93
+ Note this can be used to find token_pos but isn't used directly as
94
+ its not disciminatory enough."
95
+ (let ((name (downcase (item.name sc))))
96
+ (cond
97
+ ((string-matches name "[0-9]+")
98
+ 'numeric)
99
+ ((or (string-matches name "[0-9]+\\.[0-9]+")
100
+ (string-matches name
101
+ "[0-9][0-9]?[0-9]?,\\([0-9][0-9][0-9],\\)*[0-9][0-9][0-9]"))
102
+ 'number)
103
+ ((member_string name token_most_common)
104
+ name)
105
+ ((string-matches name ".*[^A-Za-z0-9].*")
106
+ 'sym)
107
+ ((member_string name '(jan january feb february mar march
108
+ apr april may jun june
109
+ jul july aug august sep sept september
110
+ oct october nov november dec december))
111
+ 'month)
112
+ ((member_string name '(sun sunday mon monday tue tues tuesday
113
+ wed wednesday thu thurs thursday
114
+ fri friday sat saturday))
115
+ 'day)
116
+ (t
117
+ '_other_))))
118
+
119
+ (define (token_no_starting_quote token)
120
+ "(token_no_starting_quote TOKEN)
121
+ Check to see if a single quote (or backquote) appears as prepunctuation
122
+ in this token or any previous one in this utterance. This is used to
123
+ disambiguate ending single quote as possessive or end quote."
124
+ (cond
125
+ ((null token)
126
+ t)
127
+ ((string-matches (item.feat token "prepunctuation") "[`']")
128
+ nil)
129
+ (t
130
+ (token_no_starting_quote (item.relation.prev token "Token")))))
131
+
132
+ (define (token_zerostart sc)
133
+ "(zerostart sc)
134
+ Returns, 1 if first char of sc's name is 0, 0 otherwise."
135
+ (if (string-matches (item.name sc) "^0.*")
136
+ "1"
137
+ "0"))
138
+
139
+ (define (tok_roman_to_numstring roman)
140
+ "(tok_roman_to_numstring ROMAN)
141
+ Takes a string of roman numerals and converts it to a number and
142
+ then returns the printed string of that. Only deals with numbers up to 50."
143
+ (let ((val 0) (chars (symbolexplode (upcase roman))))
144
+ (while chars
145
+ (cond
146
+ ((equal? (car chars) 'X)
147
+ (set! val (+ 10 val)))
148
+ ((equal? (car chars) 'V)
149
+ (set! val (+ 5 val)))
150
+ ((equal? (car chars) 'I)
151
+ (cond
152
+ ((equal? (car (cdr chars)) 'V)
153
+ (set! val (+ 4 val))
154
+ (set! chars (cdr chars)))
155
+ ((equal? (car (cdr chars)) 'X)
156
+ (set! val (+ 9 val))
157
+ (set! chars (cdr chars)))
158
+ (t
159
+ (set! val (+ 1 val))))))
160
+ (set! chars (cdr chars)))
161
+ (format nil "%d" val)))
162
+
163
+ (define (num_digits sc)
164
+ "(num_digits SC)
165
+ Returns number of digits (actually chars) is SC's name."
166
+ (string-length (format nil "%s" (item.name sc))))
167
+
168
+ (define (month_range sc)
169
+ "(month_range SC)
170
+ 1 if SC's name is > 0 and < 32, 0 otherwise."
171
+ (let ((val (parse-number (item.name sc))))
172
+ (if (and (> val 0) (< val 32))
173
+ "1"
174
+ "0")))
175
+
176
+ (define (remove_leading_zeros name)
177
+ "(remove_leading_zeros name)
178
+ Remove leading zeros from given string."
179
+ (let ((nname name))
180
+ (while (string-matches nname "^0..*")
181
+ (set! nname (string-after nname "0")))
182
+ nname))
183
+
184
+ (define (token_money_expand type)
185
+ "(token_money_expand type)
186
+ Convert shortened form of money identifier to words if of a known type."
187
+ (cond
188
+ ((string-equal type "HK")
189
+ (list "Hong" "Kong"))
190
+ ((string-equal type "C")
191
+ (list "Canadian"))
192
+ ((string-equal type "A")
193
+ (list "Australian"))
194
+ ((< (length type) 4)
195
+ (mapcar
196
+ (lambda (letter)
197
+ (list (list 'name letter)
198
+ (list 'pos token.letter_pos)))
199
+ (symbolexplode type)))
200
+ (t
201
+ (list type))))
202
+
203
+ (define (find_month_from_number token string-number)
204
+ "(find_month_from_number token string-number)
205
+ Find the textual representation of the month from the given string number"
206
+ (let ((nnum (parse-number string-number)))
207
+ (cond
208
+ ((equal? 1 nnum) (list "January"))
209
+ ((equal? 2 nnum) (list "February"))
210
+ ((equal? 3 nnum) (list "March"))
211
+ ((equal? 4 nnum) (list "April"))
212
+ ((equal? 5 nnum) (list "May"))
213
+ ((equal? 6 nnum) (list "June"))
214
+ ((equal? 7 nnum) (list "July"))
215
+ ((equal? 8 nnum) (list "August"))
216
+ ((equal? 9 nnum) (list "September"))
217
+ ((equal? 10 nnum) (list "October"))
218
+ ((equal? 11 nnum) (list "November"))
219
+ ((equal? 12 nnum) (list "December"))
220
+ (t
221
+ (cons "month"
222
+ (builtin_english_token_to_words token string-number))))))
223
+
224
+ (define (tok_allcaps sc)
225
+ "(tok_allcaps sc)
226
+ Returns 1 if sc's name is all capitals, 0 otherwise"
227
+ (if (string-matches (item.name sc) "[A-Z]+")
228
+ "1"
229
+ "0"))
230
+
231
+ (define (tok_section_name sc)
232
+ "(tok_section_name sc)
233
+ Returns 1 if sc's name is in list of things that are section/chapter
234
+ like."
235
+ (if (member_string
236
+ (downcase (item.name sc))
237
+ '(chapter section part article phrase verse scene act book
238
+ volume chap sect art vol war fortran saturn
239
+ trek))
240
+ "1"
241
+ "0"))
242
+
243
+ (define (tok_string_as_letters name)
244
+ "(tok_string_as_letters NAME)
245
+ Return list of letters marked as letter part of speech made
246
+ by exploding NAME."
247
+ (mapcar
248
+ (lambda (letter)
249
+ (list (list 'name letter)
250
+ (list 'pos token.letter_pos)))
251
+ (symbolexplode name)))
252
+
253
+ (define (tok_rex sc)
254
+ "(tok_rex sc)
255
+ Returns 1 if King like title is within 3 tokens before or 2 after."
256
+ (let ((kings '(king queen pope duke tsar emperor shah ceasar
257
+ duchess tsarina empress baron baroness
258
+ count countess)))
259
+ (if (or (member_string
260
+ (downcase (item.feat sc "R:Token.pp.name"))
261
+ kings)
262
+ (member_string
263
+ (downcase (item.feat sc "R:Token.pp.p.name"))
264
+ kings)
265
+ (member_string
266
+ (downcase (item.feat sc "R:Token.n.name"))
267
+ kings))
268
+ "1"
269
+ "0")))
270
+
271
+ (define (tok_rex_names sc)
272
+ "(tok_rex sc)
273
+ Returns 1 if this is a King-like name."
274
+ (if (and
275
+ (member_string
276
+ (downcase (item.name sc))
277
+ '(louis henry charles philip george edward pius william richard
278
+ ptolemy john paul peter nicholas
279
+ alexander frederick james alfonso ivan napolean leo
280
+ gregory catherine alexandria pierre elizabeth mary))
281
+ (or (string-equal "" (item.feat sc "punc"))
282
+ (string-equal "0" (item.feat sc "punc"))))
283
+ "1"
284
+ "0"))
285
+
286
+ (provide 'tokenpos)
CosyVoice-ttsfrd/resource/festival/tts.scm ADDED
@@ -0,0 +1,304 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Various tts functions and hooks
35
+
36
+ ;;; Once the utterance is built these functions synth and play it
37
+ (defvar tts_hooks (list utt.synth utt.play)
38
+ "tts_hooks
39
+ Function or list of functions to be called during text to speech.
40
+ The function tts_file, chunks data into Utterances of type Token and
41
+ applies this hook to the utterance. This typically contains the utt.synth
42
+ function and utt.play. [see TTS]")
43
+
44
+ ;;; This is used to define utterance breaks in tts on files
45
+ (defvar eou_tree
46
+ '((lisp_max_num_tokens > 200)
47
+ ((1))
48
+ ((n.whitespace matches ".*\n.*\n\\(.\\|\n\\)*");; significant break (2 nls)
49
+ ((1))
50
+ ((name matches "--+")
51
+ ((1))
52
+ ((punc matches ".*[\\?:!;].*")
53
+ ((1))
54
+ ((punc matches ".*\\..*")
55
+ ((punc matches "..+");; longer punctuation string
56
+ ((punc matches "\\..*,") ;; for U.S.S.R., like tokens
57
+ ((0))
58
+ ((1)))
59
+ ;; This is to distinguish abbreviations vs periods
60
+ ;; These are heuristics
61
+ ((name matches "\\(.*\\..*\\|[A-Z][A-Za-z]?[A-Za-z]?\\|etc\\)");; an abbreviation
62
+ ((n.whitespace is " ")
63
+ ((0));; if abbrev single space isn't enough for break
64
+ ((n.name matches "[A-Z].*")
65
+ ((1))
66
+ ((0))))
67
+ ((n.whitespace is " ");; if it doesn't look like an abbreviation
68
+ ((n.name matches "[A-Z].*");; single space and non-cap is no break
69
+ ((1))
70
+ ((0)))
71
+ ((1)))))
72
+ ((0)))))))
73
+ "eou_tree
74
+ End of utterance tree. A decision tree used to determine if the given
75
+ token marks the end of an utterance. It may look one token ahead to
76
+ do this. [see Utterance chunking]")
77
+
78
+ (define (max_num_tokens x)
79
+ "(num_tokens x)
80
+ This is probably controversial, but its good to have a maximum number
81
+ of tokens in an utterance. You really dont want to wait on very long
82
+ utterances, some utts can be thousands of words long, these maybe
83
+ shouldn't be spoken, but we do have to deal with them."
84
+ (let ((c 1) (y x))
85
+ (while y
86
+ (set! c (+ 1 c))
87
+ (set! y (item.prev y)))
88
+ c))
89
+
90
+ ;;; The program used to parse stml files
91
+ ;;; Needs version 1.0 to allow -D option to work
92
+ (defvar sgml_parse_progname "nsgmls-1.0"
93
+ "sgml_parse_progname
94
+ The name of the program to use to parse SGML files. Typically this is
95
+ nsgml-1.0 from the sp SGML package. [see XML/SGML requirements]")
96
+
97
+ ;;; When PHRASE elements are specified in an utterance in STML
98
+ ;;; no other method for phrase prediction is to be used, so we
99
+ ;;; use the following tree
100
+ (set! stml_phrase_cart_tree
101
+ '((R:Token.parent.pbreak is B)
102
+ ((B))
103
+ ((n.name is 0)
104
+ ((B))
105
+ ((NB)))))
106
+
107
+ (define (xxml_synth utt)
108
+ "(xxml_synth UTT)
109
+ This applies the xxml_hooks (mode specific) and tts_hooks to the
110
+ given utterance. This function should be called from xxml element
111
+ definitions that signal an utterance boundary."
112
+ (cond
113
+ ((or (not utt)
114
+ (not (utt.relation utt 'Token))) ;; no tokens
115
+ nil)
116
+ (t
117
+ (apply_hooks xxml_hooks utt)
118
+ (apply_hooks tts_hooks utt)
119
+ (set! utt nil) ;; not enough ...
120
+ (gc)
121
+ utt))
122
+ )
123
+
124
+ (define (xxml_attval ATTNAME ATTLIST)
125
+ "(xxml_attval ATTNAME ATTLIST)
126
+ Returns attribute value of ATTNAME in ATTLIST or nil if it doesn't
127
+ exists."
128
+ (cond
129
+ ((not ATTLIST)
130
+ nil)
131
+ ((string-equal ATTNAME (car (car ATTLIST)))
132
+ (car (cdr (car ATTLIST))))
133
+ (t
134
+ (xxml_attval ATTNAME (cdr ATTLIST)))))
135
+
136
+ (defvar xxml_word_features nil
137
+ "xxml_word_features
138
+ An assoc list of features to be added to the current word when
139
+ in xxml parse mode.")
140
+
141
+ (defvar xxml_token_hooks nil
142
+ "xxml_token_hooks
143
+ Functions to apply to each token.")
144
+
145
+ (defvar xxml_hooks nil
146
+ "xxml_hooks
147
+ Function or list of functions to be applied to an utterance when
148
+ parsed with xxML, before tts_hooks.")
149
+
150
+ (defvar xxml_elements nil
151
+ "xxml_elements
152
+ List of Scheme actions to perform on finding xxML tags.")
153
+
154
+ (defvar xml_dtd_dir libdir
155
+ "xml_dtd_dir
156
+ The directory holding standard DTD form the xml parser.")
157
+
158
+ (set! tts_fnum 1)
159
+ (define (save_tts_output utt)
160
+ (let ((fname (string-append "tts_file_" tts_fnum ".wav")))
161
+ (format stderr "festival: saving waveform in %s\n" fname)
162
+ (utt.save.wave utt fname)
163
+ (set! tts_fnum (+ 1 tts_fnum))
164
+ utt))
165
+
166
+ (define (save_waves_during_tts)
167
+ "(save_waves_during_tts)
168
+ Save each waveform in the current directory in files \"tts_file_XXX.wav\".
169
+ use (save_waves_during_tts_STOP) to stop saving waveforms"
170
+ (if (not (member save_tts_output tts_hooks))
171
+ (set! tts_hooks (append tts_hooks (list save_tts_output))))
172
+ t)
173
+
174
+ (define (save_waves_during_tts_STOP)
175
+ "(save_waves_during_tts_STOP)
176
+ Stop saving waveforms when doing tts."
177
+ (if (member save_tts_output tts_hooks)
178
+ (set! tts_hooks (delq save_tts_output tts_hooks)))
179
+ t)
180
+
181
+ (define (tts file mode)
182
+ "(tts FILE MODE)
183
+ Convert FILE to speech. MODE identifies any special treatment
184
+ necessary for FILE. This is simply a front end to tts_file but
185
+ puts the system in async audio mode first. [see TTS]"
186
+ (audio_mode 'async)
187
+ (if mode
188
+ (tts_file file mode)
189
+ (tts_file file (tts_find_text_mode file auto-text-mode-alist)))
190
+ ;; (audio_mode 'sync) ;; Hmm this is probably bad
191
+ )
192
+
193
+ (define (tts_text string mode)
194
+ "(tts_text STRING mode)
195
+ Apply tts on given string. That is, segment it into utterances and
196
+ apply tts_hooks to each utterance. This is naively done by saving the
197
+ string to a file and calling tts_file on that file. This differs from
198
+ SayText which constructs a single utterance for the whole given text."
199
+ (let ((tmpfile (make_tmp_filename))
200
+ (fd))
201
+ (set! fd (fopen tmpfile "wb"))
202
+ (format fd "%s" string)
203
+ (fclose fd)
204
+ (audio_mode 'async)
205
+ (tts_file tmpfile mode)
206
+ (delete-file tmpfile)))
207
+
208
+ (define (save_record_wave utt)
209
+ "Saves the waveform and records its so it can be joined into a
210
+ a single waveform at the end."
211
+ (let ((fn (make_tmp_filename)))
212
+ (utt.save.wave utt fn)
213
+ (set! wavefiles (cons fn wavefiles))
214
+ utt))
215
+
216
+ (define (combine_waves)
217
+ "(combine_waves)
218
+ Join all the waves together into the desired output file
219
+ and delete the intermediate ones."
220
+ (let ((wholeutt (Utterance Text "")))
221
+ (mapcar
222
+ (lambda (d)
223
+ (utt.import.wave wholeutt d t)
224
+ (delete-file d))
225
+ (reverse wavefiles))
226
+ wholeutt))
227
+
228
+ (define (tts_textall string mode)
229
+ "(tts_textall STRING MODE)
230
+ Apply tts to STRING. This function is specifically designed for
231
+ use in server mode so a single function call may synthesize the string.
232
+ This function name maybe added to the server safe functions."
233
+ (if (not (string-equal mode "nil"))
234
+ (begin
235
+ ;; a mode has been specified so do something different
236
+ (let ((tmpfile (make_tmp_filename))
237
+ (fd))
238
+ (set! fd (fopen tmpfile "wb"))
239
+ (format fd "%s" string)
240
+ (fclose fd)
241
+ (set! tts_hooks (list utt.synth save_record_wave))
242
+ (set! wavefiles nil)
243
+ (tts_file tmpfile mode)
244
+ (delete-file tmpfile)
245
+ (utt.send.wave.client (combine_waves))
246
+ ))
247
+ ;; Simple fundamental mode
248
+ (utt.send.wave.client
249
+ (utt.synth
250
+ (eval (list 'Utterance 'Text string))))))
251
+
252
+ ;; Function to interface with app_festival for asterisk
253
+ ;; See http://www.asterisk.org
254
+ (define (tts_textasterisk string mode)
255
+ "(tts_textasterisk STRING MODE)
256
+ Apply tts to STRING. This function is specifically designed for
257
+ use in server mode so a single function call may synthesize the string.
258
+ This function name may be added to the server safe functions."
259
+ (utt.send.wave.asterisk
260
+ (utt.synth
261
+ (eval (list 'Utterance 'Text string)))))
262
+
263
+
264
+
265
+ (define (tts_return_to_client)
266
+ "(tts_return_to_client)
267
+ This function is called by clients who wish to return waveforms of
268
+ their text samples asynchronously. This replaces utt.play in tts_hooks
269
+ with utt.send.wave.client."
270
+ (if (not (member utt.send.wave.client tts_hooks))
271
+ (set! tts_hooks
272
+ (append (delq utt.play tts_hooks)
273
+ (list utt.send.wave.client)))))
274
+
275
+ (defvar tts_text_modes nil
276
+ "tts_text_modes
277
+ An a-list of text modes data for file type specific tts functions.
278
+ See the manual for an example. [see Text modes]")
279
+
280
+ (define (tts_find_text_mode file alist)
281
+ "(find_text_mode FILE ALIST)
282
+ Search through ALIST for one that matches FILE. Returns nil if
283
+ nothing macthes."
284
+ (cond
285
+ ((null alist) nil) ;; can't find a match
286
+ ((string-matches file (string-append ".*" (car (car alist)) ".*"))
287
+ (cdr (car alist)))
288
+ (t
289
+ (tts_find_text_mode file (cdr alist)))))
290
+
291
+ (defvar auto-text-mode-alist
292
+ (list
293
+ (cons "\\.sable$" 'sable)
294
+ (cons "\\.ogi" 'ogimarkup)
295
+ (cons "\\.email" 'email)
296
+ (cons "" 'fundamental)
297
+ )
298
+ "auto-text-mode-alist
299
+ Following Emacs' auto-mode-alist thios provides a mechanism for auto
300
+ selecting a TTS text mode based on the filename being analyzed. Its
301
+ format is exactly the same as Emacs in that it consists of an alist of
302
+ dotted pairs of regular expression and text mode name.")
303
+
304
+ (provide 'tts)
CosyVoice-ttsfrd/resource/festival/unilex_phones.scm ADDED
@@ -0,0 +1,189 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 2003, 2004 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; unilex phoneset
35
+ ;;;
36
+
37
+
38
+ (defPhoneSet
39
+ unilex
40
+ ;;; Phone Features
41
+ (;; vowel or consonant
42
+ (vc + -)
43
+ ;; vowel length: short long dipthong schwa
44
+ (vlng s l d a 0)
45
+ ;; vowel height: high mid low
46
+ (vheight 1 2 3 0)
47
+ ;; vowel frontness: front mid back
48
+ (vfront 1 2 3 0)
49
+ ;; lip rounding
50
+ (vrnd + - 0)
51
+ ;; consonant type: stop fricative affricative nasal liquid approximant
52
+ (ctype s f a n l t r 0)
53
+ ;; place of articulation: labial alveolar palatal labio-dental
54
+ ;; dental velar glottal
55
+ (cplace l a p b d v g 0)
56
+ ;; consonant voicing
57
+ (cvox + - 0)
58
+ )
59
+ (
60
+ (SIL - 0 0 0 0 0 0 -) ;; slience ...
61
+ (# - 0 0 0 0 0 0 -) ;; slience ...
62
+ (B_10 - 0 0 0 0 0 0 -) ;; Pauses
63
+ (B_20 - 0 0 0 0 0 0 -) ;; Pauses
64
+ (B_30 - 0 0 0 0 0 0 -) ;; Pauses
65
+ (B_40 - 0 0 0 0 0 0 -) ;; Pauses
66
+ (B_50 - 0 0 0 0 0 0 -) ;; Pauses
67
+ (B_100 - 0 0 0 0 0 0 -) ;; Pauses
68
+ (B_150 - 0 0 0 0 0 0 -) ;; Pauses
69
+ (B_200 - 0 0 0 0 0 0 -) ;; Pauses
70
+ (B_250 - 0 0 0 0 0 0 -) ;; Pauses
71
+ (B_300 - 0 0 0 0 0 0 -) ;; Pauses
72
+ (B_400 - 0 0 0 0 0 0 -) ;; Pauses
73
+ (IGNORE - 0 0 0 0 0 0 -) ;; Pauses
74
+
75
+
76
+ ;; insert the phones here, see examples in
77
+ ;; festival/lib/*_phones.scm
78
+
79
+ ;(name vc vling vheight vfront vrnd ctype cplace cvox)
80
+
81
+ ;;; Rob guesed these values for Edinburgh English
82
+ ;;; Not to be taken too seriously.
83
+
84
+ (p - 0 0 0 0 s l -)
85
+ (t - 0 0 0 0 s a -)
86
+ (? - 0 0 0 0 s g +) ;;; ???
87
+ (t^ - 0 0 0 0 t a +) ;;; ???
88
+ (k - 0 0 0 0 s v -)
89
+ (x - 0 0 0 0 f v -)
90
+ (b - 0 0 0 0 s l +)
91
+ (d - 0 0 0 0 s a +)
92
+ (g - 0 0 0 0 s v +)
93
+ (ch - 0 0 0 0 a p -)
94
+ (jh - 0 0 0 0 a p +)
95
+ (s - 0 0 0 0 f a -)
96
+ (z - 0 0 0 0 f a +)
97
+ (sh - 0 0 0 0 f p -)
98
+ (zh - 0 0 0 0 f p +)
99
+ (f - 0 0 0 0 f b -)
100
+ (v - 0 0 0 0 f b +)
101
+ (th - 0 0 0 0 f d -)
102
+ (dh - 0 0 0 0 f d +)
103
+ (h - 0 0 0 0 f 0 -) ;;; ???
104
+ (m - 0 0 0 0 n l +)
105
+ (m! - 0 0 0 0 n l +)
106
+ (n - 0 0 0 0 n a +)
107
+ (n! - 0 0 0 0 n a +)
108
+ (ng - 0 0 0 0 n v +)
109
+ (l - 0 0 0 0 r a +)
110
+ (ll - 0 0 0 0 r a +)
111
+ (lw - 0 0 0 0 r a +)
112
+ (l! - 0 0 0 0 r a +)
113
+ (r - 0 0 0 0 r a +)
114
+ (y - 0 0 0 0 l p +)
115
+ (w - 0 0 0 0 l l +)
116
+ (hw - 0 0 0 0 l l +)
117
+ (e + s 2 1 - 0 0 0)
118
+ (ao + s 3 1 - 0 0 0)
119
+ (a + s 3 1 - 0 0 0)
120
+ (ah + s 3 1 - 0 0 0)
121
+ (oa + s 3 1 - 0 0 0)
122
+ (aa + s 3 1 - 0 0 0)
123
+ (ar + s 3 1 - 0 0 0)
124
+ (eh + s 3 1 - 0 0 0) ;;; ?
125
+ (oul + d 2 3 + 0 0 0) ;;; ?
126
+ (ou + d 2 3 + 0 0 0)
127
+ (ouw + d 2 3 + 0 0 0)
128
+ (oou + l 3 3 + 0 0 0)
129
+ (o + l 3 3 + 0 0 0)
130
+ (au + l 3 3 + 0 0 0)
131
+ (oo + l 3 3 + 0 0 0)
132
+ (or + l 3 3 + 0 0 0)
133
+ (our + d 2 3 + 0 0 0)
134
+ (ii + l 1 1 - 0 0 0)
135
+ (ihr + s 1 1 - 0 0 0)
136
+ (iy + l 1 1 - 0 0 0)
137
+ (i + s 1 1 - 0 0 0)
138
+ (ie + l 1 1 - 0 0 0) ;;; ?
139
+ (iii + s 1 1 - 0 0 0) ;;; was ii;
140
+ (@r + a 2 2 - r a +)
141
+ (@ + a 2 2 - 0 0 0)
142
+ (uh + s 2 2 - 0 0 0)
143
+ (uhr + s 2 2 - 0 0 0)
144
+ (u + l 1 3 + 0 0 0)
145
+ (uu + l 1 3 + 0 0 0)
146
+ (iu + l 1 3 + 0 0 0)
147
+ (uuu + l 1 3 + 0 0 0) ;;; was uu;
148
+ (uw + l 1 3 + 0 0 0) ;;; ???
149
+ (uul + l 1 3 + 0 0 0) ;;; ???
150
+ (ei + d 2 1 - 0 0 0)
151
+ (ee + d 2 1 - 0 0 0)
152
+ (ai + d 3 2 - 0 0 0) ;;; ???
153
+ (ae + d 3 2 - 0 0 0) ;;; ???
154
+ (aer + d 3 2 - 0 0 0) ;;; ???
155
+ (aai + d 3 2 - 0 0 0) ;;; ???
156
+ (oi + d 2 3 + 0 0 0) ;;; ???
157
+ (oir + d 2 3 + 0 0 0) ;;; ???
158
+ (ow + d 3 2 - 0 0 0)
159
+ (owr + d 3 2 - 0 0 0) ;;; ???
160
+ (oow + d 3 2 - 0 0 0) ;;; ???
161
+ (i@ + l 1 1 - 0 0 0) ;;; iy + @ ?
162
+ (ir + s 1 1 - 0 0 0)
163
+ (irr + s 1 1 - 0 0 0) ;;; was ir;
164
+ (iir + s 1 1 - 0 0 0)
165
+ (@@r + a 2 2 - 0 0 0)
166
+ (er + s 2 1 - 0 0 0)
167
+ (eir + s 2 1 - 0 0 0) ;;; ???
168
+ (ur + s 1 3 + 0 0 0) ;;; ???
169
+ (urr + s 1 3 + 0 0 0) ;;; ???
170
+ (iur + s 1 3 + 0 0 0) ;;; ???
171
+ )
172
+ )
173
+
174
+ (PhoneSet.silences '( # SIL))
175
+
176
+ (define (unilex::select_phoneset)
177
+ "(unilex::select_phoneset)
178
+ Set up phone set for unilex"
179
+ (Parameter.set 'PhoneSet 'unilex)
180
+ (PhoneSet.select 'unilex)
181
+ )
182
+
183
+ (define (unilex::reset_phoneset)
184
+ "(unilex::reset_phoneset)
185
+ Reset phone set for unilex."
186
+ t
187
+ )
188
+
189
+ (provide 'unilex_phones)
CosyVoice-ttsfrd/resource/festival/voices.scm ADDED
@@ -0,0 +1,361 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; Preapre to access voices. Searches down a path of places.
35
+ ;;;
36
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
37
+
38
+ (define current-voice nil
39
+ "current-voice
40
+ The name of the current voice.")
41
+
42
+ ;; The path to search for voices is created from the load-path with
43
+ ;; an extra list of directories appended.
44
+
45
+ (defvar system-voice-path '( )
46
+ "system-voice-path
47
+ Additional directory not near the load path where voices can be
48
+ found, this can be redefined in lib/sitevars.scm if desired.")
49
+
50
+ (defvar system-voice-path-multisyn '( )
51
+ "system-voice-path-multisyn
52
+ Additional directory not near the load path where multisyn voices can be
53
+ found, this can be redefined in lib/sitevars.scm if desired.")
54
+
55
+ (defvar voice-path
56
+ (remove-duplicates
57
+ (append (mapcar (lambda (d) (path-append d "voices/")) load-path)
58
+ (mapcar (lambda (d) (path-as-directory d)) system-voice-path)
59
+ ))
60
+
61
+ "voice-path
62
+ List of places to look for voices. If not set it is initialised from
63
+ load-path by appending \"voices/\" to each directory with
64
+ system-voice-path appended.")
65
+
66
+ (defvar voice-path-multisyn
67
+ (remove-duplicates
68
+ (append (mapcar (lambda (d) (path-append d "voices-multisyn/")) load-path)
69
+ (mapcar (lambda (d) (path-as-directory d)) system-voice-path-multisyn)
70
+ ))
71
+
72
+ "voice-path-multisyn
73
+ List of places to look for multisyn voices. If not set it is initialised from
74
+ load-path by appending \"voices-multisyn/\" to each directory with
75
+ system-voice-path-multisyn appended.")
76
+
77
+
78
+ ;; Declaration of voices. When we declare a voice we record the
79
+ ;; directory and set up an autoload for the vocie-selecting function
80
+
81
+ (defvar voice-locations ()
82
+ "voice-locations
83
+ Association list recording where voices were found.")
84
+
85
+ (defvar voice-location-trace nil
86
+ "voice-location-trace
87
+ Set t to print voice locations as they are found")
88
+
89
+ (define (voice-location name dir doc)
90
+ "(voice-location NAME DIR DOCSTRING)
91
+ Record the location of a voice. Called for each voice found on voice-path.
92
+ Can be called in site-init or .festivalrc for additional voices which
93
+ exist elsewhere."
94
+ (let ((func_name (intern (string-append "voice_" name)))
95
+ )
96
+
97
+ (set! name (intern name))
98
+ (set! voice-locations (cons (cons name dir) voice-locations))
99
+ (eval (list 'autoload func_name (path-append dir "festvox/" name) doc))
100
+ (if voice-location-trace
101
+ (format t "Voice: %s %s\n" name dir)
102
+ )
103
+ )
104
+ )
105
+
106
+ (define (voice-location-multisyn name rootname dir doc)
107
+ "(voice-location NAME ROOTNAME DIR DOCSTRING)
108
+ Record the location of a voice. Called for each voice found on voice-path.
109
+ Can be called in site-init or .festivalrc for additional voices which
110
+ exist elsewhere."
111
+ (let ((func_name (intern (string-append "voice_" name)))
112
+ )
113
+
114
+ (set! name (intern name))
115
+ (set! voice-locations (cons (cons name dir) voice-locations))
116
+ (eval (list 'autoload func_name (path-append dir "festvox/" rootname) doc))
117
+ (if voice-location-trace
118
+ (format t "Voice: %s %s\n" name dir)
119
+ )
120
+ )
121
+ )
122
+
123
+
124
+
125
+ (define (current_voice_reset)
126
+ "(current_voice_reset)
127
+ This function is called at the start of defining any new voice.
128
+ It is design to allow the previous voice to reset any global
129
+ values it has messed with. If this variable value is nil then
130
+ the function wont be called.")
131
+
132
+ (define (voice_reset)
133
+ "(voice_reset)
134
+ This resets all variables back to acceptable values that may affect
135
+ voice generation. This function should always be called at the
136
+ start of any function defining a voice. In addition to reseting
137
+ standard variables the function current_voice_reset will be called.
138
+ This should always be set by the voice definition function (even
139
+ if it does nothing). This allows voice specific changes to be reset
140
+ when a new voice is selection. Unfortunately I can't force this
141
+ to be used."
142
+ (Parameter.set 'Duration_Stretch 1.0)
143
+ (set! after_synth_hooks default_after_synth_hooks)
144
+
145
+ ;; The follow are reset to allow existing voices to continue
146
+ ;; to work, new voices should be setting these explicitly
147
+ (Parameter.set 'Token_Method 'Token_English)
148
+ (Parameter.set 'POS_Method Classic_POS)
149
+ (Parameter.set 'Phrasify_Method Classic_Phrasify)
150
+ (Parameter.set 'Word_Method Classic_Word)
151
+ (Parameter.set 'Pause_Method Classic_Pauses)
152
+ (Parameter.set 'PostLex_Method Classic_PostLex)
153
+
154
+ (set! diphone_module_hooks nil)
155
+ (set! UniSyn_module_hooks nil)
156
+
157
+ (if current_voice_reset
158
+ (current_voice_reset))
159
+ (set! current_voice_reset nil)
160
+ )
161
+
162
+
163
+ (defvar Voice_descriptions nil
164
+ "Internal variable containing list of voice descriptions as
165
+ decribed by proclaim_voice.")
166
+
167
+ (define (proclaim_voice name description)
168
+ "(proclaim_voice NAME DESCRIPTION)
169
+ Describe a voice to the systen. NAME should be atomic name, that
170
+ conventionally will have voice_ prepended to name the basic selection
171
+ function. OPTIONS is an assoc list of feature and value and must
172
+ have at least features for language, gender, dialect and
173
+ description. The first there of these are atomic, while the description
174
+ is a text string describing the voice."
175
+ (let ((voxdesc (assoc name Voice_descriptions)))
176
+ (if voxdesc
177
+ (set-car! (cdr voxdesc) description)
178
+ (set! Voice_descriptions
179
+ (cons (list name description) Voice_descriptions))))
180
+ )
181
+
182
+ (define (voice.description name)
183
+ "(voice.description NAME)
184
+ Output description of named voice. If the named voice is not yet loaded
185
+ it is loaded."
186
+ (let ((voxdesc (assoc name Voice_descriptions))
187
+ (cv current-voice))
188
+ (if (null voxdesc)
189
+ (unwind-protect
190
+ (begin
191
+ (voice.select name)
192
+ (voice.select cv) ;; switch back to current voice
193
+ (set! voxdesc (assoc name Voice_descriptions)))))
194
+ (if voxdesc
195
+ voxdesc
196
+ (begin
197
+ (format t "SIOD: unknown voice %s\n" name)
198
+ nil))))
199
+
200
+ (define (voice.select name)
201
+ "(voice.select NAME)
202
+ Call function to set up voice NAME. This is normally done by
203
+ prepending voice_ to NAME and call it as a function."
204
+ (eval (list (intern (string-append "voice_" name)))))
205
+
206
+ (define (voice.describe name)
207
+ "(voice.describe NAME)
208
+ Describe voice NAME by saying its description. Unfortunately although
209
+ it would be nice to say that voice's description in the voice itself
210
+ its not going to work cross language. So this just uses the current
211
+ voice. So here we assume voices describe themselves in English
212
+ which is pretty anglo-centric, shitsurei shimasu."
213
+ (let ((voxdesc (voice.description name)))
214
+ (let ((desc (car (cdr (assoc 'description (car (cdr voxdesc)))))))
215
+ (cond
216
+ (desc (tts_text desc nil))
217
+ (voxdesc
218
+ (SayText
219
+ (format nil "A voice called %s exist but it has no description"
220
+ name)))
221
+ (t
222
+ (SayText
223
+ (format nil "There is no voice called %s defined" name)))))))
224
+
225
+ (define (voice.list)
226
+ "(voice.list)
227
+ List of all (potential) voices in the system. This checks the voice-location
228
+ list of potential voices found be scanning the voice-path at start up time.
229
+ These names can be used as arguments to voice.description and
230
+ voice.describe."
231
+ (mapcar car voice-locations))
232
+
233
+ ;; Voices are found on the voice-path if they are in directories of the form
234
+ ;; DIR/LANGUAGE/NAME
235
+
236
+ (define (search-for-voices)
237
+ "(search-for-voices)
238
+ Search down voice-path to locate voices."
239
+
240
+ (let ((dirs voice-path)
241
+ (dir nil)
242
+ languages language
243
+ voices voicedir voice
244
+ )
245
+ (while dirs
246
+ (set! dir (car dirs))
247
+ (setq languages (directory-entries dir t))
248
+ (while languages
249
+ (set! language (car languages))
250
+ (set! voices (directory-entries (path-append dir language) t))
251
+ (while voices
252
+ (set! voicedir (car voices))
253
+ (set! voice (path-basename voicedir))
254
+ (if (string-matches voicedir ".*\\..*")
255
+ nil
256
+ (voice-location
257
+ voice
258
+ (path-as-directory (path-append dir language voicedir))
259
+ "voice found on path")
260
+ )
261
+ (set! voices (cdr voices))
262
+ )
263
+ (set! languages (cdr languages))
264
+ )
265
+ (set! dirs (cdr dirs))
266
+ )
267
+ )
268
+ )
269
+
270
+ ;; A single file is allowed to define multiple multisyn voices, so this has
271
+ ;; been adapted for this. Rob thinks this is just evil, but couldn't think
272
+ ;; of a better way.
273
+ (define (search-for-voices-multisyn)
274
+ "(search-for-voices-multisyn)
275
+ Search down multisyn voice-path to locate multisyn voices."
276
+ (let ((dirs voice-path-multisyn)
277
+ (dir nil)
278
+ languages language
279
+ voices voicedir voice voice-list
280
+ )
281
+ (while dirs
282
+ (set! dir (car dirs))
283
+ (set! languages (directory-entries dir t))
284
+ (while languages
285
+ (set! language (car languages))
286
+ (set! voices (directory-entries (path-append dir language) t))
287
+ (while voices
288
+ (set! voicedir (car voices))
289
+ (set! voice (path-basename voicedir))
290
+ (if (string-matches voicedir ".*\\..*")
291
+ nil
292
+ (begin
293
+ ;; load the voice definition file, but don't evaluate it!
294
+ (set! voice-def-file (load (path-append dir language voicedir "festvox"
295
+ (string-append voicedir ".scm")) t))
296
+ ;; now find the "proclaim_voice" lines and register these voices.
297
+ (mapcar
298
+ (lambda (line)
299
+ (if (string-matches (car line) "proclaim_voice")
300
+ (voice-location-multisyn (intern (cadr (cadr line))) voicedir (path-append dir language voicedir) "registerd multisyn voice")))
301
+ voice-def-file)
302
+ ))
303
+ (set! voices (cdr voices)))
304
+ (set! languages (cdr languages)))
305
+ (set! dirs (cdr dirs)))))
306
+
307
+ (search-for-voices)
308
+ (search-for-voices-multisyn)
309
+
310
+ ;; We select the default voice from a list of possibilities. One of these
311
+ ;; had better exist in every installation.
312
+
313
+ (define (no_voice_error)
314
+ (format t "\nWARNING\n")
315
+ (format t "No default voice found in %l\n" voice-path)
316
+ (format t "either no voices unpacked or voice-path is wrong\n")
317
+ (format t "Scheme interpreter will work, but there is no voice to speak with.\n")
318
+ (format t "WARNING\n\n"))
319
+
320
+ (defvar voice_default 'no_voice_error
321
+ "voice_default
322
+ A variable whose value is a function name that is called on start up to
323
+ the default voice. [see Site initialization]")
324
+
325
+ (defvar default-voice-priority-list
326
+ '(kal_diphone
327
+ cmu_us_slt_cg
328
+ cmu_us_rms_cg
329
+ cmu_us_bdl_cg
330
+ cmu_us_jmk_cg
331
+ cmu_us_awb_cg
332
+ ; cstr_rpx_nina_multisyn ; restricted license (lexicon)
333
+ ; cstr_rpx_jon_multisyn ; restricted license (lexicon)
334
+ ; cstr_edi_awb_arctic_multisyn ; restricted license (lexicon)
335
+ ; cstr_us_awb_arctic_multisyn
336
+ ked_diphone
337
+ don_diphone
338
+ rab_diphone
339
+ en1_mbrola
340
+ us1_mbrola
341
+ us2_mbrola
342
+ us3_mbrola
343
+ gsw_diphone ;; not publically distributed
344
+ el_diphone
345
+ )
346
+ "default-voice-priority-list
347
+ List of voice names. The first of them available becomes the default voice.")
348
+
349
+ (let ((voices default-voice-priority-list)
350
+ voice)
351
+ (while (and voices (eq voice_default 'no_voice_error))
352
+ (set! voice (car voices))
353
+ (if (assoc voice voice-locations)
354
+ (set! voice_default (intern (string-append "voice_" voice)))
355
+ )
356
+ (set! voices (cdr voices))
357
+ )
358
+ )
359
+
360
+
361
+ (provide 'voices)
CosyVoice-ttsfrd/resource/festival/voices/english/kal_diphone/COPYING ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ This directory contains a diphone synthesizer voice for American English.
3
+ This distribution falls under the following copyright. See further
4
+ notes below copyright.
5
+
6
+ Alan W Black and Kevin Lenzo
7
+ Copyright (c) 1998
8
+ All Rights Reserved.
9
+
10
+ Permission is hereby granted, free of charge, to use and distribute
11
+ this software and its documentation without restriction, including
12
+ without limitation the rights to use, copy, modify, merge, publish,
13
+ distribute, sublicense, and/or sell copies of this work, and to
14
+ permit persons to whom this work is furnished to do so, subject to
15
+ the following conditions:
16
+ 1. The code must retain the above copyright notice, this list of
17
+ conditions and the following disclaimer.
18
+ 2. Any modifications must be clearly marked as such.
19
+ 3. Original authors' names are not deleted.
20
+ 4. The authors' names are not used to endorse or promote products
21
+ derived from this software without specific prior written
22
+ permission.
23
+
24
+ THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK
25
+ DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING
26
+ ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT
27
+ SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE
28
+ FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
29
+ WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
30
+ AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
31
+ ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
32
+ THIS SOFTWARE.
33
+
34
+ Note this voice cotains a file Copyright CSTR, University of Edinburgh
35
+ under the same lincences as the above. The voice also requires the
36
+ separate distributions festlex_POSLEX.tar.gz and festlex_CMU.tar.gz.
37
+ festlex_POSLEX.tar.gz is copyright CSTR under the same free licence
38
+ while festlex_CMU.tar.gz contain material derived from material
39
+ copyright by Carnegie Mellon University, though their licence is as
40
+ free as the one above so.
41
+
42
+ This voice was created while awb was visiting the speech group at
43
+ CMU in fall 98.
CosyVoice-ttsfrd/resource/festival/voices/english/kal_diphone/festvox/kal_diphone.scm ADDED
@@ -0,0 +1,266 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;-*-mode:scheme-*-
2
+ ;;; ;;
3
+ ;;; Alan W Black and Kevin Lenzo ;;
4
+ ;;; Copyright (c) 1998 ;;
5
+ ;;; All Rights Reserved. ;;
6
+ ;;; ;;
7
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
8
+ ;;; this software and its documentation without restriction, including ;;
9
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
10
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
11
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
12
+ ;;; the following conditions: ;;
13
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
14
+ ;;; conditions and the following disclaimer. ;;
15
+ ;;; 2. Any modifications must be clearly marked as such. ;;
16
+ ;;; 3. Original authors' names are not deleted. ;;
17
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
18
+ ;;; derived from this software without specific prior written ;;
19
+ ;;; permission. ;;
20
+ ;;; ;;
21
+ ;;; THE AUTHORS OF THIS WORK DISCLAIM ALL WARRANTIES WITH REGARD TO ;;
22
+ ;;; THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY ;;
23
+ ;;; AND FITNESS, IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY ;;
24
+ ;;; SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
25
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
26
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
27
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
28
+ ;;; THIS SOFTWARE. ;;
29
+ ;;; ;;
30
+ ;;; This file is part "Building Voices in the Festival Speech ;;
31
+ ;;; Synthesis System" by Alan W Black and Kevin Lenzo written at ;;
32
+ ;;; Robotics Institute, Carnegie Mellon University, fall 98 ;;
33
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
34
+ ;;;
35
+ ;;; A US English diphone voice based on KAL
36
+ ;;;
37
+ ;;; Much of the front end is based on KED
38
+ ;;;
39
+
40
+ (defvar kal_diphone_dir (cdr (assoc 'kal_diphone voice-locations))
41
+ "kal_diphone_dir
42
+ The default directory for the kal diphone database.")
43
+ (set! load-path (cons (path-append kal_diphone_dir "festvox/") load-path))
44
+
45
+ (require 'radio_phones)
46
+ (require_module 'UniSyn)
47
+
48
+ ;; set this to lpc or psola
49
+ (defvar kal_sigpr 'lpc)
50
+ ;; Rset this to ungroup for ungrouped version
51
+ (defvar kal_groupungroup 'group)
52
+
53
+ (if (probe_file (path-append kal_diphone_dir "group/kallpc16k.group"))
54
+ (defvar kal_index_file
55
+ (path-append kal_diphone_dir "group/kallpc16k.group"))
56
+ (defvar kal_index_file
57
+ (path-append kal_diphone_dir "group/kallpc8k.group")))
58
+
59
+ (set! kal_psola_sep
60
+ (list
61
+ '(name "kal_psola_sep")
62
+ (list 'index_file (path-append kal_diphone_dir "dic/diphdic.est"))
63
+ '(grouped "false")
64
+ (list 'coef_dir (path-append kal_diphone_dir "pm"))
65
+ (list 'sig_dir (path-append kal_diphone_dir "wav"))
66
+ '(coef_ext ".pm")
67
+ '(sig_ext ".wav")
68
+ ; '(alternates_left ((er ax)))
69
+ ; '(alternates_right (($p p) ($k k) ($g g) ($d d) ($b b) ($t t)
70
+ ; (y ih) (ax ah) (ll l)))
71
+ '(alternates_right ((er ax)))
72
+ '(default_diphone "ax-ax")))
73
+
74
+ (set! kal_lpc_sep
75
+ (list
76
+ '(name "kal_lpc_sep")
77
+ (list 'index_file (path-append kal_diphone_dir "dic/diphdic.est"))
78
+ '(grouped "false")
79
+ (list 'coef_dir (path-append kal_diphone_dir "lpc"))
80
+ (list 'sig_dir (path-append kal_diphone_dir "lpc"))
81
+ '(coef_ext ".lpc")
82
+ '(sig_ext ".res")
83
+ '(alternates_right ((er ax)))
84
+ '(default_diphone "ax-ax")))
85
+
86
+ (set! kal_lpc_group
87
+ (list
88
+ '(name "kal_lpc_group")
89
+ (list 'index_file kal_index_file)
90
+ '(grouped "true")
91
+ '(alternates_right ((er ax)))
92
+ '(default_diphone "ax-ax")))
93
+
94
+ (defvar kal_di_16k nil)
95
+ (defvar kal_di_8k nil)
96
+
97
+ (cond
98
+ (kal_di_16k
99
+ (require 'kal_di)
100
+ (setup_kal_diphone_lpc_16k_grouped))
101
+ (kal_di_8k
102
+ (require 'kal_di)
103
+ (setup_kal_diphone_lpc_8k_grouped))
104
+ ((and (eq kal_sigpr 'psola)
105
+ (eq kal_groupungroup 'group))
106
+ (set! kal_db_name (us_diphone_init kal_psola_group)))
107
+ ((and (eq kal_sigpr 'psola)
108
+ (eq kal_groupungroup 'ungroup))
109
+ (set! kal_db_name (us_diphone_init kal_psola_sep)))
110
+ ((and (eq kal_sigpr 'lpc)
111
+ (eq kal_groupungroup 'group))
112
+ (set! kal_db_name (us_diphone_init kal_lpc_group)))
113
+ ((and (eq kal_sigpr 'lpc)
114
+ (eq kal_groupungroup 'ungroup))
115
+ (set! kal_db_name (us_diphone_init kal_lpc_sep))))
116
+
117
+ ;;;;
118
+ ;;;; Our general diphone scheme allows identification of consonant
119
+ ;;; clusters etc the follow rules should work for American English
120
+ ;;;;
121
+ (define (kal_diphone_const_clusters utt)
122
+ "(kal_diphone_const_clusters UTT)
123
+ Identify consonant clusters, dark ls etc in the segment stream
124
+ ready for diphone resynthesis. This may be called as a post lexical
125
+ rule through poslex_rule_hooks."
126
+ (mapcar
127
+ (lambda (s) (kal_diphone_fix_phone_name utt s))
128
+ (utt.relation.items utt 'Segment))
129
+ utt)
130
+
131
+ (define (kal_diphone_fix_phone_name utt seg)
132
+ "(kal_diphone_fix_phone_name UTT SEG)
133
+ Add the feature diphone_phone_name to given segment with the appropriate
134
+ name for constructing a diphone. Basically adds _ if either side is part
135
+ of the same consonant cluster, adds $ either side if in different
136
+ syllable for preceding/succeeding vowel syllable."
137
+ (let ((name (item.name seg)))
138
+ (cond
139
+ ((string-equal name "pau") t)
140
+ ((string-equal "-" (item.feat seg 'ph_vc))
141
+ (if (and (member_string name '(r w y l))
142
+ (member_string (item.feat seg "p.name") '(p t k b d g))
143
+ (item.relation.prev seg "SylStructure"))
144
+ (item.set_feat seg "us_diphone_right" (format nil "_%s" name)))
145
+ (if (and (member_string name '(w y l m n p t k))
146
+ (string-equal (item.feat seg "p.name") 's)
147
+ (item.relation.prev seg "SylStructure"))
148
+ (item.set_feat seg "us_diphone_right" (format nil "_%s" name)))
149
+ (if (and (string-equal name 's)
150
+ (member_string (item.feat seg "n.name") '(w y l m n p t k))
151
+ (item.relation.next seg "SylStructure"))
152
+ (item.set_feat seg "us_diphone_left" (format nil "%s_" name)))
153
+ (if (and (string-equal name 'hh)
154
+ (string-equal (item.feat seg "n.name") 'y)
155
+ (item.relation.next seg "SylStructure"))
156
+ (item.set_feat seg "us_diphone_left" (format nil "%s_" name)))
157
+ (if (and (string-equal name 'y)
158
+ (string-equal (item.feat seg "p.name") 'hh)
159
+ (item.relation.next seg "SylStructure"))
160
+ (item.set_feat seg "us_diphone_right" (format nil "_%s" name)))
161
+ (if (and (member_string name '(p t k b d g))
162
+ (member_string (item.feat seg "n.name") '(r w y l))
163
+ (item.relation.next seg "SylStructure"))
164
+ (item.set_feat seg "us_diphone_left" (format nil "%s_" name)))
165
+ )
166
+ ((string-equal "ah" (item.name seg))
167
+ (item.set_feat seg "us_diphone" "aa"))
168
+
169
+ )))
170
+
171
+ ;;; Set up the CMU lexicon
172
+ (setup_cmu_lex)
173
+
174
+ (define (voice_kal_diphone)
175
+ "(voice_kal_diphone)
176
+ Set up the current voice to be male American English (Kevin) using
177
+ the standard diphone code."
178
+ ;; Phone set
179
+ (voice_reset)
180
+ (Parameter.set 'Language 'americanenglish)
181
+ (require 'radio_phones)
182
+ (Parameter.set 'PhoneSet 'radio)
183
+ (PhoneSet.select 'radio)
184
+ ;; Tokenization rules
185
+ (set! token_to_words english_token_to_words)
186
+ ;; POS tagger
187
+ (require 'pos)
188
+ (set! pos_lex_name "english_poslex")
189
+ (set! pos_ngram_name 'english_pos_ngram)
190
+ (set! pos_supported t)
191
+ (set! guess_pos english_guess_pos) ;; need this for accents
192
+ ;; Lexicon selection
193
+ (lex.select "cmu")
194
+ (set! postlex_rules_hooks (list postlex_apos_s_check))
195
+ (set! postlex_s_check postlex_apos_s_check)
196
+ ;; Phrase prediction
197
+ (require 'phrase)
198
+ (Parameter.set 'Phrase_Method 'prob_models)
199
+ (set! phr_break_params english_phr_break_params)
200
+ ;; Accent and tone prediction
201
+ (require 'tobi)
202
+ (set! int_tone_cart_tree f2b_int_tone_cart_tree)
203
+ (set! int_accent_cart_tree f2b_int_accent_cart_tree)
204
+
205
+ (set! postlex_vowel_reduce_cart_tree
206
+ postlex_vowel_reduce_cart_data)
207
+ ;; F0 prediction
208
+ (require 'f2bf0lr)
209
+ (set! f0_lr_start f2b_f0_lr_start)
210
+ (set! f0_lr_mid f2b_f0_lr_mid)
211
+ (set! f0_lr_end f2b_f0_lr_end)
212
+ (Parameter.set 'Int_Method Intonation_Tree)
213
+ (set! int_lr_params
214
+ '((target_f0_mean 105) (target_f0_std 14)
215
+ (model_f0_mean 170) (model_f0_std 34)))
216
+ (Parameter.set 'Int_Target_Method Int_Targets_LR)
217
+ ;; Duration prediction
218
+ (require 'kaldurtreeZ)
219
+ (set! duration_cart_tree kal_duration_cart_tree)
220
+ (set! duration_ph_info kal_durs)
221
+ (Parameter.set 'Duration_Method Duration_Tree_ZScores)
222
+ (Parameter.set 'Duration_Stretch 1.1)
223
+ ;; Waveform synthesizer: kal diphones
224
+ ;; This assigned the diphone names from their context (_ $ etc)
225
+ (cond
226
+ ((or kal_di_16k kal_di_8k)
227
+ ;; Waveform synthesizer: kal diphones
228
+ (Parameter.set 'Synth_Method Diphone_Synthesize)
229
+ ;; This assigned the diphone names from their context (_ $ etc)
230
+ (set! diphone_module_hooks (list kal_di_const_clusters ))
231
+ (Diphone.select 'kal))
232
+ (t
233
+ (set! UniSyn_module_hooks (list kal_diphone_const_clusters ))
234
+ (set! us_abs_offset 0.0)
235
+ (set! window_factor 1.0)
236
+ (set! us_rel_offset 0.0)
237
+ (set! us_gain 0.9)
238
+
239
+ (Parameter.set 'Synth_Method 'UniSyn)
240
+ (Parameter.set 'us_sigpr kal_sigpr)
241
+ (us_db_select kal_db_name)))
242
+
243
+ (set! after_synth_hooks
244
+ (lambda (utt)
245
+ (utt.wave.rescale utt 2.6)))
246
+
247
+ (set! current-voice 'kal_diphone)
248
+ )
249
+
250
+ (proclaim_voice
251
+ 'kal_diphone
252
+ '((language english)
253
+ (gender male)
254
+ (dialect american)
255
+ (description
256
+ "This voice provides an American English male voice using a
257
+ residual excited LPC diphone synthesis method. It uses
258
+ the CMU Lexicon pronunciations. Prosodic phrasing is provided
259
+ by a statistically trained model using part of speech and local
260
+ distribution of breaks. Intonation is provided by a CART tree
261
+ predicting ToBI accents and an F0 contour generated from a model
262
+ trained from natural speech. The duration model is also trained
263
+ from data using a CART tree.")))
264
+
265
+ (provide 'kal_diphone)
266
+
CosyVoice-ttsfrd/resource/festival/voices/english/kal_diphone/festvox/kaldurtreeZ.scm ADDED
@@ -0,0 +1,978 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;;
34
+ ;;; A tree to predict zcore durations build from f2b
35
+ ;;; doesn't use actual phonemes so it can have better generalizations
36
+ ;;;
37
+ ;;; Basically copied from ked
38
+ ;;;
39
+
40
+ (set! kal_durs
41
+ '(
42
+ (uh 0.067 0.025)
43
+ (hh 0.061 0.028)
44
+ (ao 0.138 0.046)
45
+ (hv 0.053 0.020)
46
+ (v 0.051 0.019)
47
+ (ih 0.058 0.023)
48
+ (el 0.111 0.043)
49
+ (ey 0.132 0.042)
50
+ (em 0.080 0.033)
51
+ (jh 0.094 0.024)
52
+ (w 0.054 0.023)
53
+ (uw 0.107 0.044)
54
+ (ae 0.120 0.036)
55
+ (en 0.117 0.056)
56
+ (k 0.089 0.034)
57
+ (y 0.048 0.025)
58
+ (axr 0.147 0.035)
59
+ ; (l 0.056 0.026)
60
+ (l 0.066 0.026)
61
+ (ng 0.064 0.024)
62
+ (zh 0.071 0.030)
63
+ (z 0.079 0.034)
64
+ (brth 0.246 0.046)
65
+ (m 0.069 0.028)
66
+ (iy 0.097 0.041)
67
+ (n 0.059 0.025)
68
+ (ah 0.087 0.031)
69
+ (er 0.086 0.010)
70
+ (b 0.069 0.024)
71
+ (pau 0.200 0.104)
72
+ (aw 0.166 0.053)
73
+ (p 0.088 0.030)
74
+ (ch 0.115 0.025)
75
+ (ow 0.134 0.039)
76
+ (dh 0.031 0.016)
77
+ (nx 0.049 0.100)
78
+ (d 0.048 0.021)
79
+ (ax 0.046 0.024)
80
+ (h# 0.060 0.083)
81
+ (r 0.053 0.031)
82
+ (eh 0.095 0.036)
83
+ (ay 0.137 0.047)
84
+ (oy 0.183 0.050)
85
+ (f 0.095 0.033)
86
+ (sh 0.108 0.031)
87
+ (s 0.102 0.037)
88
+ (g 0.064 0.021)
89
+ (dx 0.031 0.016)
90
+ (th 0.093 0.050)
91
+ (aa 0.094 0.037)
92
+ (t 0.070 0.020)
93
+ )
94
+ )
95
+
96
+ (set! kal_duration_cart_tree
97
+ '
98
+ ((name is pau)
99
+ ((emph_sil is +)
100
+ ((0.0 -0.5))
101
+ ((p.R:SylStructure.parent.parent.pbreak is BB)
102
+ ((0.0 2.0))
103
+ ((0.0 0.0))))
104
+ ((R:SylStructure.parent.accented is 0)
105
+ ((n.ph_ctype is 0)
106
+ ((p.ph_vlng is 0)
107
+ ((R:SylStructure.parent.syl_codasize < 1.5)
108
+ ((p.ph_ctype is n)
109
+ ((ph_ctype is f)
110
+ ((0.559208 -0.783163))
111
+ ((1.05215 -0.222704)))
112
+ ((ph_ctype is s)
113
+ ((R:SylStructure.parent.syl_break is 2)
114
+ ((0.589948 0.764459))
115
+ ((R:SylStructure.parent.asyl_in < 0.7)
116
+ ((1.06385 0.567944))
117
+ ((0.691943 0.0530272))))
118
+ ((ph_vlng is l)
119
+ ((pp.ph_vfront is 1)
120
+ ((1.06991 0.766486))
121
+ ((R:SylStructure.parent.syl_break is 1)
122
+ ((0.69665 0.279248))
123
+ ((0.670353 0.0567774))))
124
+ ((p.ph_ctype is s)
125
+ ((seg_onsetcoda is coda)
126
+ ((0.828638 -0.038356))
127
+ ((ph_ctype is f)
128
+ ((0.7631 -0.545853))
129
+ ((0.49329 -0.765994))))
130
+ ((R:SylStructure.parent.parent.gpos is det)
131
+ ((R:SylStructure.parent.last_accent < 0.3)
132
+ ((R:SylStructure.parent.sub_phrases < 1)
133
+ ((0.811686 0.160195))
134
+ ((0.799015 0.713958)))
135
+ ((0.731599 -0.215472)))
136
+ ((ph_ctype is r)
137
+ ((0.673487 0.092772))
138
+ ((R:SylStructure.parent.asyl_in < 1)
139
+ ((0.745273 0.00132813))
140
+ ((0.75457 -0.334898)))))))))
141
+ ((pos_in_syl < 0.5)
142
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 2)
143
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.2)
144
+ ((0.902446 -0.041618))
145
+ ((R:SylStructure.parent.sub_phrases < 2.3)
146
+ ((0.900629 0.262952))
147
+ ((1.18474 0.594794))))
148
+ ((seg_onset_stop is 0)
149
+ ((R:SylStructure.parent.position_type is mid)
150
+ ((0.512323 -0.760444))
151
+ ((R:SylStructure.parent.syl_out < 6.8)
152
+ ((pp.ph_vlng is a)
153
+ ((0.640575 -0.450449))
154
+ ((ph_ctype is f)
155
+ ((R:SylStructure.parent.sub_phrases < 1.3)
156
+ ((0.862876 -0.296956))
157
+ ((R:SylStructure.parent.syl_out < 2.4)
158
+ ((0.803215 0.0422868))
159
+ ((0.877856 -0.154465))))
160
+ ((R:SylStructure.parent.syl_out < 3.6)
161
+ ((R:SylStructure.parent.syl_out < 1.2)
162
+ ((0.567081 -0.264199))
163
+ ((0.598043 -0.541738)))
164
+ ((0.676843 -0.166623)))))
165
+ ((0.691678 -0.57173))))
166
+ ((R:SylStructure.parent.parent.gpos is cc)
167
+ ((1.15995 0.313289))
168
+ ((pp.ph_vfront is 1)
169
+ ((0.555993 0.0695819))
170
+ ((R:SylStructure.parent.asyl_in < 1.2)
171
+ ((R:SylStructure.parent.sub_phrases < 2.7)
172
+ ((0.721635 -0.367088))
173
+ ((0.71919 -0.194887)))
174
+ ((0.547052 -0.0637491)))))))
175
+ ((ph_ctype is s)
176
+ ((R:SylStructure.parent.syl_break is 0)
177
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 1)
178
+ ((0.650007 -0.333421))
179
+ ((0.846301 -0.165383)))
180
+ ((0.527756 -0.516332)))
181
+ ((R:SylStructure.parent.syl_break is 0)
182
+ ((p.ph_ctype is s)
183
+ ((0.504414 -0.779112))
184
+ ((0.812498 -0.337611)))
185
+ ((pos_in_syl < 1.4)
186
+ ((0.513041 -0.745807))
187
+ ((p.ph_ctype is s)
188
+ ((0.350582 -1.04907))
189
+ ((0.362 -0.914974))))))))
190
+ ((R:SylStructure.parent.syl_break is 0)
191
+ ((ph_ctype is n)
192
+ ((R:SylStructure.parent.position_type is initial)
193
+ ((pos_in_syl < 1.2)
194
+ ((0.580485 0.172658))
195
+ ((0.630973 -0.101423)))
196
+ ((0.577937 -0.360092)))
197
+ ((R:SylStructure.parent.syl_out < 2.9)
198
+ ((R:SylStructure.parent.syl_out < 1.1)
199
+ ((R:SylStructure.parent.position_type is initial)
200
+ ((0.896092 0.764189))
201
+ ((R:SylStructure.parent.sub_phrases < 3.6)
202
+ ((ph_ctype is s)
203
+ ((0.877362 0.555132))
204
+ ((0.604511 0.369882)))
205
+ ((0.799982 0.666966))))
206
+ ((seg_onsetcoda is coda)
207
+ ((p.ph_vlng is a)
208
+ ((R:SylStructure.parent.last_accent < 0.4)
209
+ ((0.800736 0.240634))
210
+ ((0.720606 0.486176)))
211
+ ((1.18173 0.573811)))
212
+ ((0.607147 0.194468))))
213
+ ((ph_ctype is r)
214
+ ((0.88377 0.499383))
215
+ ((R:SylStructure.parent.last_accent < 0.5)
216
+ ((R:SylStructure.parent.position_type is initial)
217
+ ((R:SylStructure.parent.parent.word_numsyls < 2.4)
218
+ ((0.62798 0.0737318))
219
+ ((0.787334 0.331014)))
220
+ ((ph_ctype is s)
221
+ ((0.808368 0.0929299))
222
+ ((0.527948 -0.0443271))))
223
+ ((seg_coda_fric is 0)
224
+ ((p.ph_vlng is a)
225
+ ((0.679745 0.517681))
226
+ ((R:SylStructure.parent.sub_phrases < 1.1)
227
+ ((0.759979 0.128316))
228
+ ((0.775233 0.361383))))
229
+ ((R:SylStructure.parent.last_accent < 1.3)
230
+ ((0.696255 0.054136))
231
+ ((0.632425 0.246742))))))))
232
+ ((pos_in_syl < 0.3)
233
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 2)
234
+ ((0.847602 0.621547))
235
+ ((ph_ctype is s)
236
+ ((0.880645 0.501679))
237
+ ((R:SylStructure.parent.sub_phrases < 3.3)
238
+ ((R:SylStructure.parent.sub_phrases < 0.3)
239
+ ((0.901014 -0.042049))
240
+ ((0.657493 0.183226)))
241
+ ((0.680126 0.284799)))))
242
+ ((ph_ctype is s)
243
+ ((p.ph_vlng is s)
244
+ ((0.670033 -0.820934))
245
+ ((0.863306 -0.348735)))
246
+ ((ph_ctype is n)
247
+ ((R:SylStructure.parent.asyl_in < 1.2)
248
+ ((0.656966 -0.40092))
249
+ ((0.530966 -0.639366)))
250
+ ((seg_coda_fric is 0)
251
+ ((1.04153 0.364857))
252
+ ((pos_in_syl < 1.2)
253
+ ((R:SylStructure.parent.syl_out < 3.4)
254
+ ((0.81503 -0.00768613))
255
+ ((0.602665 -0.197753)))
256
+ ((0.601844 -0.394632)))))))))
257
+ ((n.ph_ctype is f)
258
+ ((pos_in_syl < 1.5)
259
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 2)
260
+ ((pos_in_syl < 0.1)
261
+ ((1.63863 0.938841))
262
+ ((R:SylStructure.parent.position_type is initial)
263
+ ((0.897722 -0.0796637))
264
+ ((nn.ph_vheight is 0)
265
+ ((0.781081 0.480026))
266
+ ((0.779711 0.127175)))))
267
+ ((ph_ctype is r)
268
+ ((p.ph_ctype is s)
269
+ ((0.581329 -0.708767))
270
+ ((0.564366 -0.236212)))
271
+ ((ph_vlng is a)
272
+ ((p.ph_ctype is r)
273
+ ((0.70992 -0.273389))
274
+ ((R:SylStructure.parent.parent.gpos is in)
275
+ ((0.764696 0.0581338))
276
+ ((nn.ph_vheight is 0)
277
+ ((0.977737 0.721904))
278
+ ((R:SylStructure.parent.sub_phrases < 2.2)
279
+ ((pp.ph_vfront is 0)
280
+ ((0.586708 0.0161206))
281
+ ((0.619949 0.227372)))
282
+ ((0.707285 0.445569))))))
283
+ ((ph_ctype is n)
284
+ ((R:SylStructure.parent.syl_break is 1)
285
+ ((nn.ph_vfront is 2)
286
+ ((0.430295 -0.120097))
287
+ ((0.741371 0.219042)))
288
+ ((0.587492 0.321245)))
289
+ ((p.ph_ctype is n)
290
+ ((0.871586 0.134075))
291
+ ((p.ph_ctype is r)
292
+ ((0.490751 -0.466418))
293
+ ((R:SylStructure.parent.syl_codasize < 1.3)
294
+ ((R:SylStructure.parent.sub_phrases < 2.2)
295
+ ((p.ph_ctype is s)
296
+ ((0.407452 -0.425925))
297
+ ((0.644771 -0.542809)))
298
+ ((0.688772 -0.201899)))
299
+ ((ph_vheight is 1)
300
+ ((nn.ph_vheight is 0)
301
+ ((0.692018 0.209018))
302
+ ((0.751345 -0.178136)))
303
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.3)
304
+ ((R:SylStructure.parent.asyl_in < 1.5)
305
+ ((0.599633 -0.235593))
306
+ ((0.60042 0.126118)))
307
+ ((p.ph_vlng is a)
308
+ ((0.7148 -0.174812))
309
+ ((R:SylStructure.parent.parent.gpos is content)
310
+ ((0.761296 -0.231509))
311
+ ((0.813081 -0.536405)))))))))))))
312
+ ((ph_ctype is n)
313
+ ((0.898844 0.163343))
314
+ ((p.ph_vlng is s)
315
+ ((seg_coda_fric is 0)
316
+ ((0.752921 -0.45528))
317
+ ((0.890079 -0.0998025)))
318
+ ((ph_ctype is f)
319
+ ((0.729376 -0.930547))
320
+ ((ph_ctype is s)
321
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 0)
322
+ ((0.745052 -0.634119))
323
+ ((0.521502 -0.760176)))
324
+ ((R:SylStructure.parent.syl_break is 1)
325
+ ((0.766575 -0.121355))
326
+ ((0.795616 -0.557509))))))))
327
+ ((p.ph_vlng is 0)
328
+ ((p.ph_ctype is r)
329
+ ((ph_vlng is 0)
330
+ ((0.733659 -0.402734))
331
+ ((R:SylStructure.parent.sub_phrases < 1.5)
332
+ ((ph_vlng is s)
333
+ ((0.326176 -0.988478))
334
+ ((n.ph_ctype is s)
335
+ ((0.276471 -0.802536))
336
+ ((0.438283 -0.900628))))
337
+ ((nn.ph_vheight is 0)
338
+ ((ph_vheight is 2)
339
+ ((0.521 -0.768992))
340
+ ((0.615436 -0.574918)))
341
+ ((ph_vheight is 1)
342
+ ((0.387376 -0.756359))
343
+ ((pos_in_syl < 0.3)
344
+ ((0.417235 -0.808937))
345
+ ((0.384043 -0.93315)))))))
346
+ ((ph_vlng is a)
347
+ ((ph_ctype is 0)
348
+ ((n.ph_ctype is s)
349
+ ((p.ph_ctype is f)
350
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.2)
351
+ ((0.415908 -0.428493))
352
+ ((pos_in_syl < 0.1)
353
+ ((0.790441 0.0211071))
354
+ ((0.452465 -0.254485))))
355
+ ((p.ph_ctype is s)
356
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.2)
357
+ ((0.582447 -0.389966))
358
+ ((0.757648 0.185781)))
359
+ ((R:SylStructure.parent.sub_phrases < 1.4)
360
+ ((0.628965 0.422551))
361
+ ((0.713613 0.145576)))))
362
+ ((seg_onset_stop is 0)
363
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 0)
364
+ ((pp.ph_vfront is 1)
365
+ ((0.412363 -0.62319))
366
+ ((R:SylStructure.parent.syl_out < 3.6)
367
+ ((0.729259 -0.317324))
368
+ ((0.441633 -0.591051))))
369
+ ((R:SylStructure.parent.syl_break is 1)
370
+ ((R:SylStructure.parent.sub_phrases < 2.7)
371
+ ((0.457728 -0.405607))
372
+ ((0.532411 -0.313148)))
373
+ ((R:SylStructure.parent.last_accent < 0.3)
374
+ ((1.14175 0.159416))
375
+ ((0.616396 -0.254651)))))
376
+ ((R:SylStructure.parent.position_type is initial)
377
+ ((0.264181 -0.799896))
378
+ ((0.439801 -0.551309)))))
379
+ ((R:SylStructure.parent.position_type is final)
380
+ ((0.552027 -0.707084))
381
+ ((0.585661 -0.901874))))
382
+ ((ph_ctype is s)
383
+ ((pos_in_syl < 1.2)
384
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.2)
385
+ ((pp.ph_vfront is 1)
386
+ ((0.607449 0.196466))
387
+ ((0.599662 0.00382414)))
388
+ ((0.64109 -0.12859)))
389
+ ((pp.ph_vfront is 1)
390
+ ((0.720484 -0.219339))
391
+ ((0.688707 -0.516734))))
392
+ ((ph_vlng is s)
393
+ ((n.ph_ctype is s)
394
+ ((R:SylStructure.parent.parent.gpos is content)
395
+ ((R:SylStructure.parent.position_type is single)
396
+ ((0.659206 0.159445))
397
+ ((R:SylStructure.parent.parent.word_numsyls < 3.5)
398
+ ((R:SylStructure.parent.sub_phrases < 2)
399
+ ((0.447186 -0.419103))
400
+ ((0.631822 -0.0928561)))
401
+ ((0.451623 -0.576116))))
402
+ ((ph_vheight is 3)
403
+ ((0.578626 -0.64583))
404
+ ((0.56636 -0.4665))))
405
+ ((R:SylStructure.parent.parent.gpos is in)
406
+ ((0.771516 -0.217292))
407
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 2)
408
+ ((0.688571 -0.304382))
409
+ ((R:SylStructure.parent.parent.gpos is content)
410
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 1)
411
+ ((n.ph_ctype is n)
412
+ ((0.556085 -0.572203))
413
+ ((0.820173 -0.240338)))
414
+ ((R:SylStructure.parent.parent.word_numsyls < 2.2)
415
+ ((0.595398 -0.588171))
416
+ ((0.524737 -0.95797))))
417
+ ((R:SylStructure.parent.sub_phrases < 3.9)
418
+ ((0.371492 -0.959427))
419
+ ((0.440479 -0.845747)))))))
420
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 0)
421
+ ((p.ph_ctype is f)
422
+ ((0.524088 -0.482247))
423
+ ((nn.ph_vheight is 1)
424
+ ((0.587666 -0.632362))
425
+ ((ph_vlng is l)
426
+ ((R:SylStructure.parent.position_type is final)
427
+ ((0.513286 -0.713117))
428
+ ((0.604613 -0.924308)))
429
+ ((R:SylStructure.parent.syl_codasize < 2.2)
430
+ ((0.577997 -0.891342))
431
+ ((0.659804 -1.15252))))))
432
+ ((pp.ph_vlng is s)
433
+ ((ph_ctype is f)
434
+ ((0.813383 -0.599624))
435
+ ((0.984027 -0.0771909)))
436
+ ((p.ph_ctype is f)
437
+ ((R:SylStructure.parent.parent.gpos is in)
438
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 1)
439
+ ((0.313572 -1.03242))
440
+ ((0.525854 -0.542799)))
441
+ ((R:SylStructure.parent.syl_out < 2.8)
442
+ ((0.613007 -0.423979))
443
+ ((0.570258 -0.766379))))
444
+ ((R:SylStructure.parent.syl_break is 1)
445
+ ((R:SylStructure.parent.parent.gpos is to)
446
+ ((0.364585 -0.792895))
447
+ ((ph_vlng is l)
448
+ ((0.69143 -0.276816))
449
+ ((0.65673 -0.523721))))
450
+ ((R:SylStructure.parent.syl_out < 3.6)
451
+ ((R:SylStructure.parent.position_type is initial)
452
+ ((0.682096 -0.488102))
453
+ ((0.406364 -0.731758)))
454
+ ((0.584694 -0.822229)))))))))))
455
+ ((n.ph_ctype is r)
456
+ ((R:SylStructure.parent.position_type is initial)
457
+ ((p.ph_vlng is a)
458
+ ((0.797058 1.02334))
459
+ ((ph_ctype is s)
460
+ ((1.0548 0.536277))
461
+ ((0.817253 0.138201))))
462
+ ((R:SylStructure.parent.sub_phrases < 1.1)
463
+ ((R:SylStructure.parent.syl_out < 3.3)
464
+ ((0.884574 -0.23471))
465
+ ((0.772063 -0.525292)))
466
+ ((nn.ph_vfront is 1)
467
+ ((1.25254 0.417485))
468
+ ((0.955557 -0.0781996)))))
469
+ ((pp.ph_vfront is 0)
470
+ ((ph_ctype is f)
471
+ ((n.ph_ctype is s)
472
+ ((R:SylStructure.parent.parent.gpos is content)
473
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 0)
474
+ ((0.583506 -0.56941))
475
+ ((0.525949 -0.289362)))
476
+ ((0.749316 -0.0921038)))
477
+ ((p.ph_vlng is s)
478
+ ((0.734234 0.139463))
479
+ ((0.680119 -0.0708717))))
480
+ ((ph_vlng is s)
481
+ ((ph_vheight is 1)
482
+ ((0.908712 -0.618971))
483
+ ((0.55344 -0.840495)))
484
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 1.2)
485
+ ((pos_in_syl < 1.2)
486
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 2)
487
+ ((0.838715 0.00913392))
488
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 1)
489
+ ((ph_vheight is 2)
490
+ ((0.555513 -0.512523))
491
+ ((R:SylStructure.parent.position_type is initial)
492
+ ((0.758711 0.121704))
493
+ ((0.737555 -0.25637))))
494
+ ((R:SylStructure.parent.syl_out < 3.1)
495
+ ((n.ph_ctype is s)
496
+ ((0.611756 -0.474522))
497
+ ((1.05437 -0.247206)))
498
+ ((R:SylStructure.parent.syl_codasize < 2.2)
499
+ ((R:SylStructure.parent.position_type is final)
500
+ ((0.567761 -0.597866))
501
+ ((0.785599 -0.407765)))
502
+ ((0.575598 -0.741256))))))
503
+ ((ph_ctype is s)
504
+ ((n.ph_ctype is s)
505
+ ((0.661069 -1.08426))
506
+ ((0.783184 -0.39789)))
507
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 1)
508
+ ((R:SylStructure.parent.sub_phrases < 2.6)
509
+ ((0.511323 -0.666011))
510
+ ((0.691878 -0.499492)))
511
+ ((ph_ctype is r)
512
+ ((0.482131 -0.253186))
513
+ ((0.852955 -0.372832))))))
514
+ ((0.854447 -0.0936489)))))
515
+ ((R:SylStructure.parent.position_type is final)
516
+ ((0.685939 -0.249982))
517
+ ((R:SylStructure.parent.syl_out < 3.2)
518
+ ((0.989843 0.18086))
519
+ ((0.686805 -0.0402908)))))))))
520
+ ((R:SylStructure.parent.syl_out < 2.4)
521
+ ((R:SylStructure.parent.syl_out < 0.2)
522
+ ((seg_onsetcoda is coda)
523
+ ((ph_ctype is s)
524
+ ((R:SylStructure.parent.syl_break is 4)
525
+ ((pp.ph_vlng is 0)
526
+ ((0.959737 1.63203))
527
+ ((1.20714 0.994933)))
528
+ ((n.ph_ctype is 0)
529
+ ((R:SylStructure.parent.syl_break is 2)
530
+ ((0.864809 0.214457))
531
+ ((0.874278 0.730381)))
532
+ ((pp.ph_vfront is 0)
533
+ ((seg_coda_fric is 0)
534
+ ((1.20844 -0.336221))
535
+ ((1.01357 0.468302)))
536
+ ((0.658106 -0.799121)))))
537
+ ((n.ph_ctype is f)
538
+ ((ph_ctype is f)
539
+ ((1.26332 0.0300613))
540
+ ((ph_vlng is d)
541
+ ((1.02719 1.1649))
542
+ ((ph_ctype is 0)
543
+ ((R:SylStructure.parent.asyl_in < 1.2)
544
+ ((1.14048 2.2668))
545
+ ((ph_vheight is 1)
546
+ ((1.15528 1.50375))
547
+ ((1.42406 2.07927))))
548
+ ((R:SylStructure.parent.sub_phrases < 1.1)
549
+ ((0.955892 1.10243))
550
+ ((R:SylStructure.parent.syl_break is 2)
551
+ ((1.32682 1.8432))
552
+ ((1.27582 1.59853)))))))
553
+ ((n.ph_ctype is 0)
554
+ ((ph_ctype is n)
555
+ ((R:SylStructure.parent.syl_break is 2)
556
+ ((1.45399 1.12927))
557
+ ((1.05543 0.442376)))
558
+ ((R:SylStructure.parent.syl_break is 4)
559
+ ((R:SylStructure.parent.position_type is final)
560
+ ((ph_ctype is f)
561
+ ((1.46434 1.76508))
562
+ ((0.978055 0.7486)))
563
+ ((1.2395 2.30826)))
564
+ ((ph_ctype is 0)
565
+ ((0.935325 1.69917))
566
+ ((nn.ph_vfront is 1)
567
+ ((1.20456 1.31128))
568
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.2)
569
+ ((nn.ph_vheight is 0)
570
+ ((1.16907 0.212421))
571
+ ((0.952091 0.653094)))
572
+ ((p.ph_ctype is 0)
573
+ ((1.05502 1.25802))
574
+ ((0.818731 0.777568))))))))
575
+ ((ph_ctype is f)
576
+ ((p.ph_ctype is 0)
577
+ ((1.03918 0.163941))
578
+ ((0.737545 -0.167063)))
579
+ ((R:SylStructure.parent.position_type is final)
580
+ ((n.ph_ctype is n)
581
+ ((R:SylStructure.parent.last_accent < 0.5)
582
+ ((R:SylStructure.parent.sub_phrases < 2.8)
583
+ ((0.826207 -0.000859005))
584
+ ((0.871119 0.273433)))
585
+ ((R:SylStructure.parent.parent.word_numsyls < 2.4)
586
+ ((1.17405 1.05694))
587
+ ((0.858394 0.244916))))
588
+ ((R:SylStructure.parent.syl_codasize < 2.2)
589
+ ((p.ph_ctype is 0)
590
+ ((1.14092 1.21187))
591
+ ((R:SylStructure.parent.syl_break is 2)
592
+ ((1.02653 0.59865))
593
+ ((0.94248 1.1634))))
594
+ ((seg_coda_fric is 0)
595
+ ((1.07441 0.292935))
596
+ ((1.15736 0.92574)))))
597
+ ((ph_vlng is s)
598
+ ((R:SylStructure.parent.syl_break is 2)
599
+ ((1.34638 1.23484))
600
+ ((0.951514 2.02008)))
601
+ ((ph_ctype is 0)
602
+ ((p.ph_ctype is r)
603
+ ((0.806106 0.697089))
604
+ ((R:SylStructure.parent.syl_break is 2)
605
+ ((1.10891 0.992197))
606
+ ((1.04657 1.51093))))
607
+ ((1.18165 0.520952)))))))))
608
+ ((p.ph_vlng is 0)
609
+ ((pos_in_syl < 0.7)
610
+ ((R:SylStructure.parent.position_type is final)
611
+ ((ph_ctype is r)
612
+ ((0.966357 0.185827))
613
+ ((ph_ctype is s)
614
+ ((0.647163 0.0332298))
615
+ ((0.692972 -0.534917))))
616
+ ((ph_ctype is s)
617
+ ((0.881521 0.575107))
618
+ ((p.ph_ctype is f)
619
+ ((0.8223 -0.111275))
620
+ ((R:SylStructure.parent.last_accent < 0.3)
621
+ ((0.969188 0.09447))
622
+ ((0.894438 0.381947))))))
623
+ ((p.ph_ctype is f)
624
+ ((0.479748 -0.490108))
625
+ ((0.813125 -0.201268))))
626
+ ((ph_ctype is s)
627
+ ((0.908566 1.20397))
628
+ ((R:SylStructure.parent.last_accent < 1.2)
629
+ ((0.88078 0.636568))
630
+ ((0.978087 1.07763))))))
631
+ ((pos_in_syl < 1.3)
632
+ ((R:SylStructure.parent.syl_break is 0)
633
+ ((pos_in_syl < 0.1)
634
+ ((R:SylStructure.parent.position_type is initial)
635
+ ((p.ph_ctype is n)
636
+ ((0.801651 -0.0163359))
637
+ ((ph_ctype is s)
638
+ ((n.ph_ctype is r)
639
+ ((0.893307 1.07253))
640
+ ((p.ph_vlng is 0)
641
+ ((0.92651 0.525806))
642
+ ((0.652444 0.952792))))
643
+ ((p.ph_vlng is 0)
644
+ ((seg_onsetcoda is coda)
645
+ ((0.820151 0.469117))
646
+ ((p.ph_ctype is f)
647
+ ((0.747972 -0.0716448))
648
+ ((ph_ctype is f)
649
+ ((0.770882 0.457137))
650
+ ((0.840905 0.102492)))))
651
+ ((R:SylStructure.parent.syl_out < 1.1)
652
+ ((0.667824 0.697337))
653
+ ((0.737967 0.375114))))))
654
+ ((ph_vheight is 1)
655
+ ((0.624353 0.410671))
656
+ ((R:SylStructure.parent.asyl_in < 0.8)
657
+ ((0.647905 -0.331055))
658
+ ((p.ph_ctype is s)
659
+ ((0.629039 -0.240616))
660
+ ((0.749277 -0.0191273))))))
661
+ ((ph_vheight is 3)
662
+ ((p.ph_ctype is s)
663
+ ((0.626922 0.556537))
664
+ ((0.789357 0.153892)))
665
+ ((seg_onsetcoda is coda)
666
+ ((n.ph_ctype is 0)
667
+ ((R:SylStructure.parent.parent.word_numsyls < 3.4)
668
+ ((0.744714 0.123242))
669
+ ((0.742039 0.295753)))
670
+ ((seg_coda_fric is 0)
671
+ ((R:SylStructure.parent.parent.word_numsyls < 2.4)
672
+ ((ph_vheight is 1)
673
+ ((0.549715 -0.341018))
674
+ ((0.573641 -0.00893114)))
675
+ ((nn.ph_vfront is 2)
676
+ ((0.67099 -0.744625))
677
+ ((0.664438 -0.302803))))
678
+ ((p.ph_vlng is 0)
679
+ ((0.630028 0.113815))
680
+ ((0.632794 -0.128733)))))
681
+ ((ph_ctype is r)
682
+ ((0.367169 -0.854509))
683
+ ((0.94334 -0.216179))))))
684
+ ((n.ph_ctype is f)
685
+ ((ph_vlng is 0)
686
+ ((1.3089 0.46195))
687
+ ((R:SylStructure.parent.syl_codasize < 1.3)
688
+ ((1.07673 0.657169))
689
+ ((pp.ph_vlng is 0)
690
+ ((0.972319 1.08222))
691
+ ((1.00038 1.46257)))))
692
+ ((p.ph_vlng is l)
693
+ ((1.03617 0.785204))
694
+ ((p.ph_vlng is a)
695
+ ((R:SylStructure.parent.position_type is final)
696
+ ((1.00681 0.321168))
697
+ ((0.928115 0.950834)))
698
+ ((ph_vlng is 0)
699
+ ((pos_in_syl < 0.1)
700
+ ((R:SylStructure.parent.position_type is final)
701
+ ((0.863682 -0.167374))
702
+ ((nn.ph_vheight is 0)
703
+ ((p.ph_ctype is f)
704
+ ((0.773591 -0.00374425))
705
+ ((R:SylStructure.parent.syl_out < 1.1)
706
+ ((0.951802 0.228448))
707
+ ((1.02282 0.504252))))
708
+ ((1.09721 0.736476))))
709
+ ((R:SylStructure.parent.position_type is final)
710
+ ((1.04302 0.0590974))
711
+ ((0.589208 -0.431535))))
712
+ ((n.ph_ctype is 0)
713
+ ((1.27879 1.00642))
714
+ ((ph_vlng is s)
715
+ ((R:SylStructure.parent.asyl_in < 1.4)
716
+ ((0.935787 0.481652))
717
+ ((0.9887 0.749861)))
718
+ ((R:SylStructure.parent.syl_out < 1.1)
719
+ ((R:SylStructure.parent.position_type is final)
720
+ ((0.921307 0.0696307))
721
+ ((0.83675 0.552212)))
722
+ ((0.810076 -0.0479225))))))))))
723
+ ((ph_ctype is s)
724
+ ((n.ph_ctype is s)
725
+ ((0.706959 -1.0609))
726
+ ((p.ph_ctype is n)
727
+ ((0.850614 -0.59933))
728
+ ((n.ph_ctype is r)
729
+ ((0.665947 0.00698725))
730
+ ((n.ph_ctype is 0)
731
+ ((R:SylStructure.parent.position_type is initial)
732
+ ((0.762889 -0.0649044))
733
+ ((0.723956 -0.248899)))
734
+ ((R:SylStructure.parent.sub_phrases < 1.4)
735
+ ((0.632957 -0.601987))
736
+ ((0.889114 -0.302401)))))))
737
+ ((ph_ctype is f)
738
+ ((R:SylStructure.parent.syl_codasize < 2.2)
739
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.2)
740
+ ((R:SylStructure.parent.syl_out < 1.1)
741
+ ((0.865267 0.164636))
742
+ ((0.581827 -0.0989051)))
743
+ ((nn.ph_vfront is 2)
744
+ ((0.684459 -0.316836))
745
+ ((0.778854 -0.0961191))))
746
+ ((R:SylStructure.parent.syl_out < 1.1)
747
+ ((p.ph_ctype is s)
748
+ ((0.837964 -0.429437))
749
+ ((0.875304 -0.0652743)))
750
+ ((0.611071 -0.635089))))
751
+ ((p.ph_ctype is r)
752
+ ((R:SylStructure.parent.syl_out < 1.1)
753
+ ((0.762012 0.0139361))
754
+ ((0.567983 -0.454845)))
755
+ ((R:SylStructure.parent.syl_codasize < 2.2)
756
+ ((ph_ctype is l)
757
+ ((1.18845 0.809091))
758
+ ((R:SylStructure.parent.position_type is initial)
759
+ ((ph_ctype is n)
760
+ ((0.773548 -0.277092))
761
+ ((1.01586 0.281001)))
762
+ ((p.ph_ctype is 0)
763
+ ((1.06831 0.699145))
764
+ ((0.924189 0.241873)))))
765
+ ((R:SylStructure.parent.syl_break is 0)
766
+ ((ph_ctype is n)
767
+ ((0.592321 -0.470784))
768
+ ((0.778688 -0.072112)))
769
+ ((n.ph_ctype is s)
770
+ ((1.08848 0.0733489))
771
+ ((1.25674 0.608371))))))))))
772
+ ((pos_in_syl < 0.7)
773
+ ((p.ph_vlng is 0)
774
+ ((R:SylStructure.parent.position_type is mid)
775
+ ((ph_ctype is 0)
776
+ ((ph_vheight is 2)
777
+ ((0.456225 -0.293282))
778
+ ((0.561529 -0.0816115)))
779
+ ((0.6537 -0.504024)))
780
+ ((ph_ctype is s)
781
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 2)
782
+ ((1.31586 0.98395))
783
+ ((R:SylStructure.parent.position_type is single)
784
+ ((0.816869 0.634789))
785
+ ((R:SylStructure.parent.syl_out < 4.4)
786
+ ((1.05578 0.479029))
787
+ ((R:SylStructure.parent.asyl_in < 0.4)
788
+ ((1.11813 0.143214))
789
+ ((0.87178 0.406834))))))
790
+ ((n.ph_ctype is n)
791
+ ((R:SylStructure.parent.last_accent < 0.6)
792
+ ((0.838154 -0.415599))
793
+ ((0.924024 0.110288)))
794
+ ((seg_onsetcoda is coda)
795
+ ((nn.ph_vfront is 2)
796
+ ((0.670096 0.0314187))
797
+ ((n.ph_ctype is f)
798
+ ((1.00363 0.693893))
799
+ ((R:SylStructure.parent.syl_out < 6)
800
+ ((0.772363 0.215675))
801
+ ((0.920313 0.574068)))))
802
+ ((R:SylStructure.parent.position_type is final)
803
+ ((0.673837 -0.458142))
804
+ ((R:SylStructure.parent.sub_phrases < 2.8)
805
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 2)
806
+ ((0.894817 0.304628))
807
+ ((ph_ctype is n)
808
+ ((0.787302 -0.23094))
809
+ ((R:SylStructure.parent.asyl_in < 1.2)
810
+ ((ph_ctype is f)
811
+ ((R:SylStructure.parent.last_accent < 0.5)
812
+ ((1.12278 0.326954))
813
+ ((0.802236 -0.100616)))
814
+ ((0.791255 -0.0919132)))
815
+ ((0.95233 0.219053)))))
816
+ ((R:SylStructure.parent.position_type is initial)
817
+ ((ph_ctype is f)
818
+ ((1.0616 0.216118))
819
+ ((0.703216 -0.00834086)))
820
+ ((ph_ctype is f)
821
+ ((1.22277 0.761763))
822
+ ((0.904811 0.332721))))))))))
823
+ ((ph_vheight is 0)
824
+ ((p.ph_vlng is s)
825
+ ((0.873379 0.217178))
826
+ ((n.ph_ctype is r)
827
+ ((0.723915 1.29451))
828
+ ((n.ph_ctype is 0)
829
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 1)
830
+ ((R:SylStructure.parent.sub_phrases < 4)
831
+ ((seg_coda_fric is 0)
832
+ ((p.ph_vlng is l)
833
+ ((0.849154 0.945261))
834
+ ((0.633261 0.687498)))
835
+ ((0.728546 0.403076)))
836
+ ((0.850962 1.00255)))
837
+ ((0.957999 1.09113)))
838
+ ((0.85771 0.209045)))))
839
+ ((ph_vheight is 2)
840
+ ((0.803401 -0.0544067))
841
+ ((0.681353 0.256045)))))
842
+ ((n.ph_ctype is f)
843
+ ((ph_ctype is s)
844
+ ((p.ph_vlng is 0)
845
+ ((0.479307 -0.9673))
846
+ ((0.700477 -0.351397)))
847
+ ((ph_ctype is f)
848
+ ((0.73467 -0.6233))
849
+ ((R:SylStructure.parent.syl_break is 0)
850
+ ((p.ph_ctype is s)
851
+ ((0.56282 0.266234))
852
+ ((p.ph_ctype is r)
853
+ ((0.446203 -0.302281))
854
+ ((R:SylStructure.parent.sub_phrases < 2.7)
855
+ ((ph_ctype is 0)
856
+ ((0.572016 -0.0102436))
857
+ ((0.497358 -0.274514)))
858
+ ((0.545477 0.0482177)))))
859
+ ((ph_vlng is s)
860
+ ((0.805269 0.888495))
861
+ ((ph_ctype is n)
862
+ ((0.869854 0.653018))
863
+ ((R:SylStructure.parent.sub_phrases < 2.2)
864
+ ((0.735031 0.0612886))
865
+ ((0.771859 0.346637))))))))
866
+ ((R:SylStructure.parent.syl_codasize < 1.4)
867
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.3)
868
+ ((R:SylStructure.parent.position_type is initial)
869
+ ((0.743458 0.0411808))
870
+ ((1.13068 0.613305)))
871
+ ((pos_in_syl < 1.2)
872
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 1)
873
+ ((1.11481 0.175467))
874
+ ((0.937893 -0.276407)))
875
+ ((0.74264 -0.550878))))
876
+ ((pos_in_syl < 3.4)
877
+ ((seg_onsetcoda is coda)
878
+ ((ph_ctype is r)
879
+ ((n.ph_ctype is s)
880
+ ((0.714319 -0.240328))
881
+ ((p.ph_ctype is 0)
882
+ ((0.976987 0.330352))
883
+ ((1.1781 -0.0816682))))
884
+ ((ph_ctype is l)
885
+ ((n.ph_ctype is 0)
886
+ ((1.39137 0.383533))
887
+ ((0.725585 -0.324515)))
888
+ ((ph_vheight is 3)
889
+ ((ph_vlng is d)
890
+ ((0.802626 -0.62487))
891
+ ((n.ph_ctype is r)
892
+ ((0.661091 -0.513869))
893
+ ((R:SylStructure.parent.position_type is initial)
894
+ ((R:SylStructure.parent.parent.word_numsyls < 2.4)
895
+ ((0.482285 0.207874))
896
+ ((0.401601 -0.0204711)))
897
+ ((0.733755 0.397372)))))
898
+ ((n.ph_ctype is r)
899
+ ((p.ph_ctype is 0)
900
+ ((pos_in_syl < 1.2)
901
+ ((0.666325 0.271734))
902
+ ((nn.ph_vheight is 0)
903
+ ((0.642401 -0.261466))
904
+ ((0.783684 -0.00956571))))
905
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.2)
906
+ ((0.692225 -0.381895))
907
+ ((0.741921 -0.0898767))))
908
+ ((nn.ph_vfront is 2)
909
+ ((ph_ctype is s)
910
+ ((0.697527 -1.12626))
911
+ ((n.ph_ctype is s)
912
+ ((ph_vlng is 0)
913
+ ((R:SylStructure.parent.sub_phrases < 2.4)
914
+ ((0.498719 -0.906926))
915
+ ((0.635342 -0.625651)))
916
+ ((0.45886 -0.385089)))
917
+ ((0.848596 -0.359702))))
918
+ ((p.ph_vlng is a)
919
+ ((p.ph_ctype is 0)
920
+ ((0.947278 0.216904))
921
+ ((0.637933 -0.394349)))
922
+ ((p.ph_ctype is r)
923
+ ((R:SylStructure.parent.syl_break is 0)
924
+ ((0.529903 -0.860573))
925
+ ((0.581378 -0.510488)))
926
+ ((ph_vlng is 0)
927
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.2)
928
+ ((seg_onset_stop is 0)
929
+ ((R:SylStructure.parent.syl_break is 0)
930
+ ((p.ph_vlng is d)
931
+ ((0.768363 0.0108428))
932
+ ((ph_ctype is s)
933
+ ((0.835756 -0.035054))
934
+ ((ph_ctype is f)
935
+ ((p.ph_vlng is s)
936
+ ((0.602016 -0.179727))
937
+ ((0.640126 -0.297341)))
938
+ ((0.674628 -0.542602)))))
939
+ ((ph_ctype is s)
940
+ ((0.662261 -0.60496))
941
+ ((0.662088 -0.432058))))
942
+ ((R:SylStructure.parent.syl_out < 4.4)
943
+ ((0.582448 -0.389079))
944
+ ((ph_ctype is s)
945
+ ((0.60413 -0.73564))
946
+ ((0.567153 -0.605444)))))
947
+ ((R:SylStructure.parent.R:Syllable.p.syl_break is 2)
948
+ ((0.761115 -0.827377))
949
+ ((ph_ctype is n)
950
+ ((0.855183 -0.275338))
951
+ ((R:SylStructure.parent.syl_break is 0)
952
+ ((0.788288 -0.802801))
953
+ ((R:SylStructure.parent.syl_codasize < 2.2)
954
+ ((0.686134 -0.371234))
955
+ ((0.840184 -0.772883)))))))
956
+ ((pos_in_syl < 1.2)
957
+ ((R:SylStructure.parent.syl_break is 0)
958
+ ((n.ph_ctype is n)
959
+ ((0.423592 -0.655006))
960
+ ((R:SylStructure.parent.syl_out < 4.4)
961
+ ((0.595269 -0.303751))
962
+ ((0.478433 -0.456882))))
963
+ ((0.688133 -0.133182)))
964
+ ((seg_onset_stop is 0)
965
+ ((1.27464 0.114442))
966
+ ((0.406837 -0.167545))))))))))))
967
+ ((ph_ctype is r)
968
+ ((0.462874 -0.87695))
969
+ ((R:SylStructure.parent.R:Syllable.n.syl_onsetsize < 0.2)
970
+ ((0.645442 -0.640572))
971
+ ((0.673717 -0.321322)))))
972
+ ((0.61008 -0.925472))))))))
973
+ ;; RMSE 0.8085 Correlation is 0.5899 Mean (abs) Error 0.6024 (0.5393)
974
+
975
+
976
+ ))
977
+
978
+ (provide 'kaldurtreeZ)
CosyVoice-ttsfrd/resource/festival/voices/english/kal_diphone/group/kallpc16k.group ADDED
The diff for this file is too large to render. See raw diff
 
CosyVoice-ttsfrd/resource/festival/voices/english/rab_diphone/COPYING ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ This directory contains a diphone synthesizer voice for British English.
3
+ This distribution falls under the following copyright. See further
4
+ notes below copyright.
5
+
6
+ Centre for Speech Technology Research
7
+ University of Edinburgh, UK
8
+ Copyright (c) 1996,1997
9
+ All Rights Reserved.
10
+
11
+ Permission is hereby granted, free of charge, to use and distribute
12
+ this software and its documentation without restriction, including
13
+ without limitation the rights to use, copy, modify, merge, publish,
14
+ distribute, sublicense, and/or sell copies of this work, and to
15
+ permit persons to whom this work is furnished to do so, subject to
16
+ the following conditions:
17
+ 1. The code must retain the above copyright notice, this list of
18
+ conditions and the following disclaimer.
19
+ 2. Any modifications must be clearly marked as such.
20
+ 3. Original authors' names are not deleted.
21
+ 4. The authors' names are not used to endorse or promote products
22
+ derived from this software without specific prior written
23
+ permission.
24
+
25
+ THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK
26
+ DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING
27
+ ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT
28
+ SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE
29
+ FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
30
+ WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
31
+ AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
32
+ ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF
33
+ THIS SOFTWARE.
34
+
35
+ Note this voice (as it currently is set up) also requires the separate
36
+ distributions festlex_POSLEX.tar.gz and festlex_OALD.tar.gz.
37
+ festlex_POSLEX.tar.gz fall under the same copyright as the above
38
+ but festlex_OALD.tar.gz is free for non-commercial use only.
CosyVoice-ttsfrd/resource/festival/voices/english/rab_diphone/festvox/rab_diphone.scm ADDED
@@ -0,0 +1,263 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2
+ ;;; ;;
3
+ ;;; Centre for Speech Technology Research ;;
4
+ ;;; University of Edinburgh, UK ;;
5
+ ;;; Copyright (c) 1996,1997 ;;
6
+ ;;; All Rights Reserved. ;;
7
+ ;;; ;;
8
+ ;;; Permission is hereby granted, free of charge, to use and distribute ;;
9
+ ;;; this software and its documentation without restriction, including ;;
10
+ ;;; without limitation the rights to use, copy, modify, merge, publish, ;;
11
+ ;;; distribute, sublicense, and/or sell copies of this work, and to ;;
12
+ ;;; permit persons to whom this work is furnished to do so, subject to ;;
13
+ ;;; the following conditions: ;;
14
+ ;;; 1. The code must retain the above copyright notice, this list of ;;
15
+ ;;; conditions and the following disclaimer. ;;
16
+ ;;; 2. Any modifications must be clearly marked as such. ;;
17
+ ;;; 3. Original authors' names are not deleted. ;;
18
+ ;;; 4. The authors' names are not used to endorse or promote products ;;
19
+ ;;; derived from this software without specific prior written ;;
20
+ ;;; permission. ;;
21
+ ;;; ;;
22
+ ;;; THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK ;;
23
+ ;;; DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ;;
24
+ ;;; ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT ;;
25
+ ;;; SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE ;;
26
+ ;;; FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES ;;
27
+ ;;; WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN ;;
28
+ ;;; AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ;;
29
+ ;;; ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF ;;
30
+ ;;; THIS SOFTWARE. ;;
31
+ ;;; ;;
32
+ ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
33
+ ;;; Set up rab_diphones using the standard UniSyn diphone synthesizer
34
+ ;;;
35
+ ;;; Roger diphones: male RP English collected October 1996
36
+ ;;;
37
+
38
+ (defvar rab_diphone_dir (cdr (assoc 'rab_diphone voice-locations))
39
+ "rab_diphone_dir
40
+ The default directory for the rab diphone database.")
41
+
42
+ (require 'mrpa_phones)
43
+ (require 'pos)
44
+ (require 'phrase)
45
+ (require 'tobi)
46
+ (require 'f2bf0lr)
47
+ (require 'mrpa_durs)
48
+ (require 'gswdurtreeZ)
49
+ (require_module 'UniSyn)
50
+
51
+ (setup_oald_lex)
52
+ ;; set this to lpc or psola
53
+ (defvar rab_sigpr 'lpc)
54
+ ;; Rset this to ungroup for ungrouped version
55
+ (defvar rab_groupungroup 'group)
56
+
57
+ (if (probe_file (path-append rab_diphone_dir "group/rablpc16k.group"))
58
+ (defvar rab_index_file
59
+ (path-append rab_diphone_dir "group/rablpc16k.group"))
60
+ (defvar rab_index_file
61
+ (path-append rab_diphone_dir "group/rablpc8k.group")))
62
+
63
+ (set! rab_psola_sep
64
+ (list
65
+ '(name "rab_psola_sep")
66
+ (list 'index_file (path-append rab_diphone_dir "dic/diphdic_full.est"))
67
+ '(grouped "false")
68
+ (list 'coef_dir (path-append rab_diphone_dir "pm"))
69
+ (list 'sig_dir (path-append rab_diphone_dir "wav"))
70
+ '(coef_ext ".pm")
71
+ '(sig_ext ".wav")))
72
+
73
+ (set! rab_lpc_sep
74
+ (list
75
+ '(name "rab_lpc_sep")
76
+ (list 'index_file (path-append rab_diphone_dir "dic/diphdic_full.est"))
77
+ '(grouped "false")
78
+ (list 'coef_dir (path-append rab_diphone_dir "lpc"))
79
+ (list 'sig_dir (path-append rab_diphone_dir "lpc"))
80
+ '(coef_ext ".lpc")
81
+ '(sig_ext ".res")))
82
+
83
+ (set! rab_psola_group
84
+ (list
85
+ '(name "rab_psola_group")
86
+ (list 'index_file
87
+ (path-append rab_diphone_dir "group/rab.group"))
88
+ '(grouped "true")))
89
+
90
+ (set! rab_lpc_group
91
+ (list
92
+ '(name "rab_lpc_group")
93
+ (list 'index_file rab_index_file)
94
+ '(alternates_left ((i ii) (ll l) (u uu) (i@ ii) (uh @) (a aa)
95
+ (u@ uu) (w @) (o oo) (e@ ei) (e ei)
96
+ (r @)))
97
+ '(alternates_right ((i ii) (ll l) (u uu) (i@ ii)
98
+ (y i) (uh @) (r @) (w @)))
99
+ '(default_diphone @-@@)
100
+ '(grouped "true")))
101
+
102
+ ;;; Setup the desried DB
103
+ (cond
104
+ ((and (eq rab_sigpr 'psola)
105
+ (eq rab_groupungroup 'group))
106
+ (set! rab_db_name (us_diphone_init rab_psola_group)))
107
+ ((and (eq rab_sigpr 'psola)
108
+ (eq rab_groupungroup 'ungroup))
109
+ (set! rab_db_name (us_diphone_init rab_psola_sep)))
110
+ ((and (eq rab_sigpr 'lpc)
111
+ (eq rab_groupungroup 'group))
112
+ (set! rab_db_name (us_diphone_init rab_lpc_group)))
113
+ ((and (eq rab_sigpr 'lpc)
114
+ (eq rab_groupungroup 'ungroup))
115
+ (set! rab_db_name (us_diphone_init rab_lpc_sep))))
116
+
117
+ (define (rab_postlex_syllabics utt)
118
+ "(rab_postlex_syllabics utt)
119
+ Because the lexicon is somewhat random in its used of syllable l n and
120
+ m this is designed to post process the output inserting schwa before
121
+ them. Ideally the lexicon should be fixed."
122
+ (mapcar
123
+ (lambda (s)
124
+ (if (and (member_string (item.name s) '("l" "n" "m"))
125
+ (string-equal "coda" (item.feat s "seg_onsetcoda"))
126
+ (not (member_string (item.feat s "p.name") '(l r)))
127
+ (string-equal "-" (item.feat s "p.ph_vc")))
128
+ (item.relation.insert
129
+ s 'SylStructure
130
+ (item.insert s (list "@") 'before)
131
+ 'before)))
132
+ (utt.relation.items utt 'Segment)))
133
+
134
+ (define (rab_diphone_const_clusters utt)
135
+ "(rab_diphone_const_clusters UTT)
136
+ Identify consonant clusters, dark ls etc in the segment item
137
+ ready for diphone resynthesis. This may be called as a post lexical
138
+ rule through poslex_rule_hooks."
139
+ (mapcar
140
+ (lambda (s) (rab_diphone_fix_phone_name utt s))
141
+ (utt.relation.items utt 'Segment))
142
+ utt)
143
+
144
+ (define (rab_diphone_fix_phone_name utt seg)
145
+ "(rab_diphone_fix_phone_name UTT SEG)
146
+ Add the feature diphone_phone_name to given segment with the appropriate
147
+ name for constructing a diphone. Basically adds _ if either side is part
148
+ of the same consonant cluster, adds $ either side if in different
149
+ syllable for preceding/succeeding vowel syllable, and converts l to ll
150
+ in coda part of syllables."
151
+ (let ((name (item.name seg)))
152
+ (cond
153
+ ((string-equal name "#") t)
154
+ ((string-equal "-" (item.feat seg 'ph_vc))
155
+ (if (and (member_string name '(r w y l))
156
+ (member_string (item.feat seg "p.name") '(p t k b d g))
157
+ (item.relation.prev seg "SylStructure"))
158
+ (item.set_feat seg "us_diphone_right" (format nil "_%s" name)))
159
+ (if (and (member_string name '(w y l m n p t k))
160
+ (string-equal (item.feat seg "p.name") 's)
161
+ (item.relation.prev seg "SylStructure"))
162
+ (item.set_feat seg "us_diphone_right" (format nil "_%s" name)))
163
+ (if (and (string-equal name 's)
164
+ (member_string (item.feat seg "n.name") '(w y l m n p t k))
165
+ (item.relation.next seg "SylStructure"))
166
+ (item.set_feat seg "us_diphone_left" (format nil "%s_" name)))
167
+ (if (and (member_string name '(p t k b d g))
168
+ (member_string (item.feat seg "n.name") '(r w y l))
169
+ (item.relation.next seg "SylStructure"))
170
+ (item.set_feat seg "us_diphone_left" (format nil "%s_" name)))
171
+ (if (and (member_string name '(p k b d g))
172
+ (string-equal "+" (item.feat seg 'p.ph_vc))
173
+ (not (member_string (item.feat seg "p.name") '(@ aa o)))
174
+ (not (item.relation.prev seg "SylStructure")))
175
+ (item.set_feat seg "us_diphone_right" (format nil "$%s" name)))
176
+ (if (and (member_string name '(p t k b d g))
177
+ (string-equal "+" (item.feat seg 'n.ph_vc))
178
+ (not (member_string (item.feat seg "n.name") '(@ aa)))
179
+ (not (item.relation.next seg "SylStructure")))
180
+ (item.set_feat seg "us_diphone_left" (format nil "%s$" name)))
181
+ (if (and (string-equal "l" name)
182
+ (string-equal "+" (item.feat seg "p.ph_vc"))
183
+ (not (string-equal "a" (item.feat seg "p.ph_vlng")))
184
+ (item.relation.prev seg 'SylStructure))
185
+ (item.set_feat seg "us_diphone_right" "ll"))
186
+ (if (and (member_string name '(ch jh))
187
+ (string-equal "+" (item.feat seg 'p.ph_vc)))
188
+ (item.set_feat seg "us_diphone_right" "t"))
189
+ )
190
+ )))
191
+
192
+ (define (voice_rab_diphone)
193
+ "(voice_rab_diphone)
194
+ Set up the current voice to be a British male RP (Roger) speaker using
195
+ the rab diphone set."
196
+ (voice_reset)
197
+ (Parameter.set 'Language 'britishenglish)
198
+ ;; Phone set
199
+ (Parameter.set 'PhoneSet 'mrpa)
200
+ (PhoneSet.select 'mrpa)
201
+ ;; Tokenization rules
202
+ (set! token_to_words english_token_to_words)
203
+ ;; POS tagger
204
+ (set! pos_lex_name "english_poslex")
205
+ (set! pos_ngram_name 'english_pos_ngram)
206
+ (set! pos_supported t)
207
+ (set! guess_pos english_guess_pos) ;; need this for accents
208
+ ;; Lexicon selection
209
+ (lex.select "oald")
210
+ (set! postlex_rules_hooks (list postlex_apos_s_check
211
+ rab_postlex_syllabics))
212
+ (set! postlex_s_check postlex_apos_s_check)
213
+ ;; Phrase prediction
214
+ (Parameter.set 'Phrase_Method 'prob_models)
215
+ (set! phr_break_params english_phr_break_params)
216
+ ;; Accent and tone prediction
217
+ (set! int_tone_cart_tree f2b_int_tone_cart_tree)
218
+ (set! int_accent_cart_tree f2b_int_accent_cart_tree)
219
+ ;; F0 prediction
220
+ (set! f0_lr_start f2b_f0_lr_start)
221
+ (set! f0_lr_mid f2b_f0_lr_mid)
222
+ (set! f0_lr_end f2b_f0_lr_end)
223
+ (Parameter.set 'Int_Method Intonation_Tree)
224
+ (set! int_lr_params
225
+ '((target_f0_mean 105) (target_f0_std 14)
226
+ (model_f0_mean 170) (model_f0_std 34)))
227
+ (Parameter.set 'Int_Target_Method Int_Targets_LR)
228
+ ;; Duration prediction -- use gsw durations
229
+ (set! duration_cart_tree gsw_duration_cart_tree)
230
+ (set! duration_ph_info gsw_durs)
231
+ (Parameter.set 'Duration_Method Duration_Tree_ZScores)
232
+ (Parameter.set 'Duration_Stretch 1.05)
233
+ ;; Waveform synthesizer: Roger diphones
234
+ ;; This assigned the diphone names from their context (_ $ etc)
235
+ (set! UniSyn_module_hooks (list rab_diphone_const_clusters ))
236
+ (set! us_abs_offset 0.0)
237
+ (set! window_factor 1.0)
238
+ (set! us_rel_offset 0.0)
239
+ (set! us_gain 0.9)
240
+
241
+ (Parameter.set 'Synth_Method 'UniSyn)
242
+ (Parameter.set 'us_sigpr rab_sigpr)
243
+ (us_db_select rab_db_name)
244
+
245
+ (set! current-voice 'rab_diphone)
246
+ )
247
+
248
+ (proclaim_voice
249
+ 'rab_diphone
250
+ '((language english)
251
+ (gender male)
252
+ (dialect british)
253
+ (description
254
+ "This voice provides a British RP English male voice using a
255
+ residual excited LPC diphone synthesis method. It uses a
256
+ modified Oxford Advanced Learners' Dictionary for pronunciations.
257
+ Prosodic phrasing is provided by a statistically trained model
258
+ using part of speech and local distribution of breaks. Intonation
259
+ is provided by a CART tree predicting ToBI accents and an F0
260
+ contour generated from a model trained from natural speech. The
261
+ duration model is also trained from data using a CART tree.")))
262
+
263
+ (provide 'rab_diphone)
CosyVoice-ttsfrd/resource/festival/voices/english/rab_diphone/group/rablpc16k.group ADDED
The diff for this file is too large to render. See raw diff