longtc commited on
Commit
ca83dc0
·
verified ·
1 Parent(s): 8a383bb

Upload 3 files

Browse files
Files changed (3) hide show
  1. README +145 -0
  2. metadata.csv +0 -0
  3. wavs.zip +3 -0
README ADDED
@@ -0,0 +1,145 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ -----------------------------------------------------------------------------
2
+ The LJ Speech Dataset
3
+
4
+ Version 1.0
5
+ July 5, 2017
6
+ https://keithito.com/LJ-Speech-Dataset
7
+ -----------------------------------------------------------------------------
8
+
9
+
10
+ OVERVIEW
11
+
12
+ This is a public domain speech dataset consisting of 13,100 short audio clips
13
+ of a single speaker reading passages from 7 non-fiction books. A transcription
14
+ is provided for each clip. Clips vary in length from 1 to 10 seconds and have
15
+ a total length of approximately 24 hours.
16
+
17
+ The texts were published between 1884 and 1964, and are in the public domain.
18
+ The audio was recorded in 2016-17 by the LibriVox project and is also in the
19
+ public domain.
20
+
21
+
22
+
23
+ FILE FORMAT
24
+
25
+ Metadata is provided in metadata.csv. This file consists of one record per
26
+ line, delimited by the pipe character (0x7c). The fields are:
27
+
28
+ 1. ID: this is the name of the corresponding .wav file
29
+ 2. Transcription: words spoken by the reader (UTF-8)
30
+ 3. Normalized Transcription: transcription with numbers, ordinals, and
31
+ monetary units expanded into full words (UTF-8).
32
+
33
+ Each audio file is a single-channel 16-bit PCM WAV with a sample rate of
34
+ 22050 Hz.
35
+
36
+
37
+
38
+ STATISTICS
39
+
40
+ Total Clips 13,100
41
+ Total Words 225,715
42
+ Total Characters 1,308,674
43
+ Total Duration 23:55:17
44
+ Mean Clip Duration 6.57 sec
45
+ Min Clip Duration 1.11 sec
46
+ Max Clip Duration 10.10 sec
47
+ Mean Words per Clip 17.23
48
+ Distinct Words 13,821
49
+
50
+
51
+
52
+ MISCELLANEOUS
53
+
54
+ The audio clips range in length from approximately 1 second to 10 seconds.
55
+ They were segmented automatically based on silences in the recording. Clip
56
+ boundaries generally align with sentence or clause boundaries, but not always.
57
+
58
+ The text was matched to the audio manually, and a QA pass was done to ensure
59
+ that the text accurately matched the words spoken in the audio.
60
+
61
+ The original LibriVox recordings were distributed as 128 kbps MP3 files. As a
62
+ result, they may contain artifacts introduced by the MP3 encoding.
63
+
64
+ The following abbreviations appear in the text. They may be expanded as
65
+ follows:
66
+
67
+ Abbreviation Expansion
68
+ --------------------------
69
+ Mr. Mister
70
+ Mrs. Misess (*)
71
+ Dr. Doctor
72
+ No. Number
73
+ St. Saint
74
+ Co. Company
75
+ Jr. Junior
76
+ Maj. Major
77
+ Gen. General
78
+ Drs. Doctors
79
+ Rev. Reverend
80
+ Lt. Lieutenant
81
+ Hon. Honorable
82
+ Sgt. Sergeant
83
+ Capt. Captain
84
+ Esq. Esquire
85
+ Ltd. Limited
86
+ Col. Colonel
87
+ Ft. Fort
88
+
89
+ * there's no standard expansion of "Mrs."
90
+
91
+
92
+ 19 of the transcriptions contain non-ASCII characters (for example, LJ016-0257
93
+ contains "raison d'être").
94
+
95
+ For more information or to report errors, please email [email protected].
96
+
97
+
98
+
99
+ LICENSE
100
+
101
+ This dataset is in the public domain in the USA (and likely other countries as
102
+ well). There are no restrictions on its use. For more information, please see:
103
+ https://librivox.org/pages/public-domain.
104
+
105
+
106
+ CHANGELOG
107
+
108
+ * 1.0 (July 8, 2017):
109
+ Initial release
110
+
111
+ * 1.1 (Feb 19, 2018):
112
+ Version 1.0 included 30 .wav files with no corresponding annotations in
113
+ metadata.csv. These have been removed in version 1.1. Thanks to Rafael Valle
114
+ for spotting this.
115
+
116
+
117
+ CREDITS
118
+
119
+ This dataset consists of excerpts from the following works:
120
+
121
+ * Morris, William, et al. Arts and Crafts Essays. 1893.
122
+ * Griffiths, Arthur. The Chronicles of Newgate, Vol. 2. 1884.
123
+ * Roosevelt, Franklin D. The Fireside Chats of Franklin Delano Roosevelt.
124
+ 1933-42.
125
+ * Harland, Marion. Marion Harland's Cookery for Beginners. 1893.
126
+ * Rolt-Wheeler, Francis. The Science - History of the Universe, Vol. 5:
127
+ Biology. 1910.
128
+ * Banks, Edgar J. The Seven Wonders of the Ancient World. 1916.
129
+ * President's Commission on the Assassination of President Kennedy. Report
130
+ of the President's Commission on the Assassination of President Kennedy.
131
+ 1964.
132
+
133
+ Recordings by Linda Johnson. Alignment and annotation by Keith Ito. All text,
134
+ audio, and annotations are in the public domain.
135
+
136
+ There's no requirement to cite this work, but if you'd like to do so, you can
137
+ link to: https://keithito.com/LJ-Speech-Dataset
138
+
139
+ or use the following:
140
+ @misc{ljspeech17,
141
+ author = {Keith Ito},
142
+ title = {The LJ Speech Dataset},
143
+ howpublished = {\url{https://keithito.com/LJ-Speech-Dataset/}},
144
+ year = 2017
145
+ }
metadata.csv ADDED
The diff for this file is too large to render. See raw diff
 
wavs.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6308ac6b78142ea4deb5c0e3d06d2da593942ad2b7bd5320afde121250501d23
3
+ size 3184114031