File size: 10,685 Bytes
92674ed
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;-*-mode:scheme-*-
;;;                                                                       ;;
;;;                    Alan W Black and Kevin Lenzo                       ;;
;;;                         Copyright (c) 1998                            ;;
;;;                        All Rights Reserved.                           ;;
;;;                                                                       ;;
;;;  Permission is hereby granted, free of charge, to use and distribute  ;;
;;;  this software and its documentation without restriction, including   ;;
;;;  without limitation the rights to use, copy, modify, merge, publish,  ;;
;;;  distribute, sublicense, and/or sell copies of this work, and to      ;;
;;;  permit persons to whom this work is furnished to do so, subject to   ;;
;;;  the following conditions:                                            ;;
;;;   1. The code must retain the above copyright notice, this list of    ;;
;;;      conditions and the following disclaimer.                         ;;
;;;   2. Any modifications must be clearly marked as such.                ;;
;;;   3. Original authors' names are not deleted.                         ;;
;;;   4. The authors' names are not used to endorse or promote products   ;;
;;;      derived from this software without specific prior written        ;;
;;;      permission.                                                      ;;
;;;                                                                       ;;
;;;  THE AUTHORS OF THIS WORK DISCLAIM ALL WARRANTIES WITH REGARD TO      ;;
;;;  THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY   ;;
;;;  AND FITNESS, IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY         ;;
;;;  SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES            ;;
;;;  WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN   ;;
;;;  AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,          ;;
;;;  ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF       ;;
;;;  THIS SOFTWARE.                                                       ;;
;;;                                                                       ;;
;;;  This file is part "Building Voices in the Festival Speech            ;;
;;;  Synthesis System" by Alan W Black and Kevin Lenzo written at         ;;
;;;  Robotics Institute, Carnegie Mellon University, fall 98              ;;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;;;
;;;  A US English diphone voice based on KAL
;;;
;;;  Much of the front end is based on KED
;;;

(defvar kal_diphone_dir (cdr (assoc 'kal_diphone voice-locations))
  "kal_diphone_dir
  The default directory for the kal diphone database.")
(set! load-path (cons (path-append kal_diphone_dir "festvox/") load-path))

(require 'radio_phones)
(require_module 'UniSyn)

;; set this to lpc or psola
(defvar kal_sigpr 'lpc)
;; Rset this to ungroup for ungrouped version
(defvar kal_groupungroup 'group)

(if (probe_file (path-append kal_diphone_dir "group/kallpc16k.group"))
    (defvar kal_index_file 
      (path-append kal_diphone_dir "group/kallpc16k.group"))
    (defvar kal_index_file 
      (path-append kal_diphone_dir "group/kallpc8k.group")))

(set! kal_psola_sep 
      (list
       '(name "kal_psola_sep")
       (list 'index_file (path-append kal_diphone_dir "dic/diphdic.est"))
       '(grouped "false")
       (list 'coef_dir (path-append kal_diphone_dir "pm"))
       (list 'sig_dir (path-append kal_diphone_dir "wav"))
       '(coef_ext ".pm")
       '(sig_ext ".wav")
;       '(alternates_left ((er ax)))
;       '(alternates_right (($p p) ($k k) ($g g) ($d d) ($b b) ($t t)
;				  (y ih) (ax ah) (ll l)))
       '(alternates_right ((er ax)))
       '(default_diphone "ax-ax")))

(set! kal_lpc_sep 
      (list
       '(name "kal_lpc_sep")
       (list 'index_file (path-append kal_diphone_dir "dic/diphdic.est"))
       '(grouped "false")
       (list 'coef_dir (path-append kal_diphone_dir "lpc"))
       (list 'sig_dir  (path-append kal_diphone_dir "lpc"))
       '(coef_ext ".lpc")
       '(sig_ext ".res")
       '(alternates_right ((er ax)))
       '(default_diphone "ax-ax")))

(set! kal_lpc_group 
      (list
       '(name "kal_lpc_group")
       (list 'index_file kal_index_file)
       '(grouped "true")
       '(alternates_right ((er ax)))
       '(default_diphone "ax-ax")))

(defvar kal_di_16k nil)
(defvar kal_di_8k nil)

(cond
 (kal_di_16k
  (require 'kal_di)
  (setup_kal_diphone_lpc_16k_grouped))
 (kal_di_8k
  (require 'kal_di)
  (setup_kal_diphone_lpc_8k_grouped))
 ((and (eq kal_sigpr 'psola)
       (eq kal_groupungroup 'group))
  (set! kal_db_name (us_diphone_init kal_psola_group)))
 ((and (eq kal_sigpr 'psola)
       (eq kal_groupungroup 'ungroup))
  (set! kal_db_name (us_diphone_init kal_psola_sep)))
 ((and (eq kal_sigpr 'lpc)
       (eq kal_groupungroup 'group))
  (set! kal_db_name (us_diphone_init kal_lpc_group)))
 ((and (eq kal_sigpr 'lpc)
       (eq kal_groupungroup 'ungroup))
  (set! kal_db_name (us_diphone_init kal_lpc_sep))))

;;;;
;;;;  Our general diphone scheme allows identification of consonant
;;;   clusters etc the follow rules should work for American English
;;;;
(define (kal_diphone_const_clusters utt)
"(kal_diphone_const_clusters UTT)
Identify consonant clusters, dark ls etc in the segment stream
ready for diphone resynthesis.  This may be called as a post lexical
rule through poslex_rule_hooks."
  (mapcar
   (lambda (s) (kal_diphone_fix_phone_name utt s))
   (utt.relation.items utt 'Segment))
  utt)

(define (kal_diphone_fix_phone_name utt seg)
"(kal_diphone_fix_phone_name UTT SEG)
Add the feature diphone_phone_name to given segment with the appropriate
name for constructing a diphone.  Basically adds _ if either side is part
of the same consonant cluster, adds $ either side if in different
syllable for preceding/succeeding vowel syllable."
  (let ((name (item.name seg)))
    (cond
     ((string-equal name "pau") t)
     ((string-equal "-" (item.feat seg 'ph_vc))
      (if (and (member_string name '(r w y l))
	       (member_string (item.feat seg "p.name") '(p t k b d g))
	       (item.relation.prev seg "SylStructure"))
	  (item.set_feat seg "us_diphone_right" (format nil "_%s" name)))
      (if (and (member_string name '(w y l m n p t k))
	       (string-equal (item.feat seg "p.name") 's)
	       (item.relation.prev seg "SylStructure"))
	  (item.set_feat seg "us_diphone_right" (format nil "_%s" name)))
      (if (and (string-equal name 's)
	       (member_string (item.feat seg "n.name") '(w y l m n p t k))
	       (item.relation.next seg "SylStructure"))
	  (item.set_feat seg "us_diphone_left" (format nil "%s_" name)))
      (if (and (string-equal name 'hh)
	       (string-equal (item.feat seg "n.name") 'y)
	       (item.relation.next seg "SylStructure"))
	  (item.set_feat seg "us_diphone_left" (format nil "%s_" name)))
      (if (and (string-equal name 'y)
	       (string-equal (item.feat seg "p.name") 'hh)
	       (item.relation.next seg "SylStructure"))
	  (item.set_feat seg "us_diphone_right" (format nil "_%s" name)))
      (if (and (member_string name '(p t k b d g))
	       (member_string (item.feat seg "n.name") '(r w y l))
	       (item.relation.next seg "SylStructure"))
	  (item.set_feat seg "us_diphone_left" (format nil "%s_" name)))
      )
     ((string-equal "ah" (item.name seg))
      (item.set_feat seg "us_diphone" "aa"))

   )))

;;;  Set up the CMU lexicon
(setup_cmu_lex)

(define (voice_kal_diphone)
"(voice_kal_diphone)
 Set up the current voice to be male  American English (Kevin) using
 the standard diphone code."
  ;; Phone set
  (voice_reset)
  (Parameter.set 'Language 'americanenglish)
  (require 'radio_phones)
  (Parameter.set 'PhoneSet 'radio)
  (PhoneSet.select 'radio)
  ;; Tokenization rules
  (set! token_to_words english_token_to_words)
  ;; POS tagger
  (require 'pos)
  (set! pos_lex_name "english_poslex")
  (set! pos_ngram_name 'english_pos_ngram)
  (set! pos_supported t)
  (set! guess_pos english_guess_pos)   ;; need this for accents
  ;; Lexicon selection
  (lex.select "cmu")
  (set! postlex_rules_hooks (list postlex_apos_s_check))
  (set! postlex_s_check postlex_apos_s_check)
  ;; Phrase prediction
  (require 'phrase)
  (Parameter.set 'Phrase_Method 'prob_models)
  (set! phr_break_params english_phr_break_params)
  ;; Accent and tone prediction
  (require 'tobi)
  (set! int_tone_cart_tree f2b_int_tone_cart_tree)
  (set! int_accent_cart_tree f2b_int_accent_cart_tree)

  (set! postlex_vowel_reduce_cart_tree 
	postlex_vowel_reduce_cart_data)
  ;; F0 prediction
  (require 'f2bf0lr)
  (set! f0_lr_start f2b_f0_lr_start)
  (set! f0_lr_mid f2b_f0_lr_mid)
  (set! f0_lr_end f2b_f0_lr_end)
  (Parameter.set 'Int_Method Intonation_Tree)
  (set! int_lr_params
	'((target_f0_mean 105) (target_f0_std 14)
	  (model_f0_mean 170) (model_f0_std 34)))
  (Parameter.set 'Int_Target_Method Int_Targets_LR)
  ;; Duration prediction
  (require 'kaldurtreeZ)
  (set! duration_cart_tree kal_duration_cart_tree)
  (set! duration_ph_info kal_durs)
  (Parameter.set 'Duration_Method Duration_Tree_ZScores)
  (Parameter.set 'Duration_Stretch 1.1)
  ;; Waveform synthesizer: kal diphones
  ;; This assigned the diphone names from their context (_ $ etc)
  (cond
   ((or kal_di_16k kal_di_8k)
    ;; Waveform synthesizer: kal diphones
    (Parameter.set 'Synth_Method Diphone_Synthesize)
    ;; This assigned the diphone names from their context (_ $ etc)
    (set! diphone_module_hooks (list kal_di_const_clusters ))
    (Diphone.select 'kal))
   (t
    (set! UniSyn_module_hooks (list kal_diphone_const_clusters ))
    (set! us_abs_offset 0.0)
    (set! window_factor 1.0)
    (set! us_rel_offset 0.0)
    (set! us_gain 0.9)

    (Parameter.set 'Synth_Method 'UniSyn)
    (Parameter.set 'us_sigpr kal_sigpr)
    (us_db_select kal_db_name)))

  (set! after_synth_hooks 
	(lambda (utt) 
	  (utt.wave.rescale utt 2.6)))

  (set! current-voice 'kal_diphone)
)

(proclaim_voice
 'kal_diphone
 '((language english)
   (gender male)
   (dialect american)
   (description
    "This voice provides an American English male voice using a
     residual excited LPC diphone synthesis method.  It uses 
     the CMU Lexicon pronunciations.  Prosodic phrasing is provided 
     by a statistically trained model using part of speech and local 
     distribution of breaks.  Intonation is provided by a CART tree 
     predicting ToBI accents and an F0 contour generated from a model 
     trained from natural speech.  The duration model is also trained 
     from data using a CART tree.")))

(provide 'kal_diphone)