carmentano commited on
Commit
a3227a0
verified
1 Parent(s): 46428ff

Upload README.md

Browse files
Files changed (1) hide show
  1. README.md +327 -1
README.md CHANGED
@@ -1,3 +1,329 @@
 
1
  ---
2
- license: cc-by-4.0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
  ---
3
+ annotations_creators:
4
+ - expert-generated
5
+ language:
6
+ - ca
7
+ language_creators:
8
+ - found
9
+ license:
10
+ - cc-by-4.0
11
+ multilinguality:
12
+ - monolingual
13
+ pretty_name: |-
14
+ annotated_catalan_common_voice_v17
15
+ size_categories:
16
+ - 1M<n<10M
17
+ source_datasets:
18
+ - extended|common_voice
19
+ tags: []
20
+ task_categories:
21
+ - automatic-speech-recognition
22
+ - audio-classification
23
+ - audio-to-audio
24
+ task_ids:
25
+ - audio-language-identification
26
  ---
27
+
28
+ # Catalan Common Voice v17 - metadata annotated version
29
+
30
+ ## Table of Contents
31
+ - [Dataset Card Creation Guide](#dataset-card-creation-guide)
32
+ - [Table of Contents](#table-of-contents)
33
+ - [Dataset Description](#dataset-description)
34
+ - [Dataset Summary](#dataset-summary)
35
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
36
+ - [Languages](#languages)
37
+ - [Dataset Structure](#dataset-structure)
38
+ - [Data Instances](#data-instances)
39
+ - [Data Fields](#data-fields)
40
+ - [Data Splits](#data-splits)
41
+ - [Dataset Creation](#dataset-creation)
42
+ - [Curation Rationale](#curation-rationale)
43
+ - [Source Data](#source-data)
44
+ - [Initial Data Collection and Normalization](#initial-data-collection-and-normalization)
45
+ - [Who are the source language producers?](#who-are-the-source-language-producers)
46
+ - [Annotations](#annotations)
47
+ - [Annotation process](#annotation-process)
48
+ - [Who are the annotators?](#who-are-the-annotators)
49
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
50
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
51
+ - [Social Impact of Dataset](#social-impact-of-dataset)
52
+ - [Discussion of Biases](#discussion-of-biases)
53
+ - [Other Known Limitations](#other-known-limitations)
54
+ - [Additional Information](#additional-information)
55
+ - [Dataset Curators](#dataset-curators)
56
+ - [Licensing Information](#licensing-information)
57
+ - [Citation Information](#citation-information)
58
+ - [Contributions](#contributions)
59
+
60
+ ## Dataset Description
61
+
62
+ - **Homepage:** [https://projecteaina.cat/tech/](https://projecteaina.cat/tech/)
63
+ - **Point of Contact:** [langech@bsc.es](langech@bsc.es)
64
+
65
+ ### Dataset Summary
66
+
67
+ This version of the [Catalan sentences of the Common Voice corpus (v17)](https://commonvoice.mozilla.org/en/datasets)
68
+ includes metadata (gender and accent) for 263 speakers annotated by a team of experts.
69
+
70
+ They correspond to the speakers who have recorded more than 1200 seconds of Speech in [v13](https://huggingface.co/datasets/mozilla-foundation/common_voice_13_0/tree/main/transcript/ca).
71
+
72
+ We release these annotations (``annotated_accent`` and ``annotated_gender`` fields)
73
+ along with information about the agreement among the annotators (``annotated_accent_agreement`` and ``annotated_gender_agreement``).
74
+
75
+ The overall quality of the recordings has also been annotated for these speakers (``mean_quality`` and ``stdev_quality`` fields).
76
+
77
+ Additionally, since for certain users self-declared information appears only in some of their recordings (``client_id`` field),
78
+ we have created new columns with metadata extracted from other recordings of the same user (``propagated_gender`` and ``propagated_accents`` fields).
79
+
80
+ To facilitate data analysis and classification tasks,
81
+ we have normalized the information in the free-form ``accents`` field to the predefined list used until version 8 of the corpus (``propagated_accents_norm``).
82
+ For each user, we have added the fields ``assigned_accent`` and ``assigned_gender``.
83
+ For speakers annotated by the team of experts, this information is derived from their annotations.
84
+ For other speakers, we have extracted this information from their self-declared data.
85
+ See [annotations](#annotations) for more details.
86
+
87
+ ### Supported Tasks and Leaderboards
88
+
89
+ Automatic Speech Recognition, Gender classification, Accent classification.
90
+
91
+ ### Languages
92
+
93
+ The dataset is in Catalan (ca).
94
+
95
+ ## Dataset Structure
96
+
97
+ ### Data Instances
98
+
99
+
100
+ ```
101
+ {
102
+ 'client_id': '69dafb41ddc0ea2785719305fdc5c8d79c4b2829d9f3325bda707dcaa553f95c5fbf4b072970d9004d3e31543fcb2c55e252dc904c4fb5aee2a5e5500df90967',
103
+ 'path': 'common_voice_ca_19909748.mp3',
104
+ 'sentence': 'En el carrer de l'esgl茅sia es troben bona part dels edificis importants de la vila.',
105
+ 'up_votes': 2,
106
+ 'down_votes': 0,
107
+ 'age': 'thirties',
108
+ 'gender': 'male',
109
+ 'accent': 'balear',
110
+ 'variant': '',
111
+ 'locale': 'ca',
112
+ 'segment': '',
113
+ 'mean quality': '4.0',
114
+ 'stdev quality': '0.0',
115
+ 'annotated_accent': 'balearic',
116
+ 'annotated_accent_agreement': '100.0',
117
+ 'annotated_gender': 'male_masculine',
118
+ 'annotated_gender_agreement': '100.0',
119
+ 'propagated_gender': 'male_masculine',
120
+ 'propagated_accents': 'balear',
121
+ 'propagated_accents_normalized': 'balearic',
122
+ 'assigned_accent': 'balearic',
123
+ 'assigned_gender': 'male'
124
+
125
+ }
126
+
127
+ ```
128
+
129
+
130
+ ### Data Fields
131
+
132
+ Most of the data fields come from the original [Common Voice corpus](https://commonvoice.mozilla.org/en/datasets):
133
+
134
+ - `client_id`(string): An id for which client (voice) made the recording
135
+
136
+ - `path`(string): The path to the audio file
137
+
138
+ - `sentence_id`(string): An id for the text sentence
139
+
140
+ - `sentence`(string): The sentence the user was prompted to speak
141
+
142
+ - `sentence_domain`(string): Semantic domain of the sentence
143
+
144
+ - `up_votes`(int64): How many upvotes the audio file has received from reviewers
145
+
146
+ - `down_votes`(int64): How many downvotes the audio file has received from reviewers
147
+
148
+ - `age`(string): Self-reported age of the speaker (e.g. teens, twenties, fifties)
149
+
150
+ - `gender`(string): Self-reported gender of the speaker
151
+
152
+ - `accent`(string): Self-reported accent of the speaker
153
+
154
+ - `locale`(string): The locale of the speaker
155
+
156
+ - `segment`(string): Usually an empty field
157
+
158
+ In this version of the corpus, we have added the following fields:
159
+
160
+ - `annotated_gender`(string): Annotated gender by the experts team.
161
+
162
+ - `annotated_gender_agreement` (float): Agreement whithin the annotation team about the gender of the speaker.
163
+
164
+ - `annotated_accent` (string): Annotated accent by the experts team. The accents considered are: Balearic, Central, Northern, Northwestern, Valencian.
165
+
166
+ - `annotated_accent_agreement` (float): Agreement whithin the annotaion team about the accent of the speaker.
167
+
168
+ - `mean quality` (float): Mean annotated quality of the speakers' recording.
169
+
170
+ - `stdev quality` (float): Deviation in the quality annotation between annotators.
171
+
172
+ - `propagated_gender` (string): Self-declared gender as indicated in certain recordings by the user. Speakers that change self-declared gender have been labeled as "other.
173
+
174
+ - `propagated_accents` (string): Self-declared accent as indicated in certain recordings by the user. See [annotations](#annotations) for more information.
175
+
176
+ - `propagated_accents_normalized` (string): Propagated accent, normalized to the closed-options list used until version 7.
177
+
178
+ - `assigned_accent` (string): Accent assigned to the speaker.
179
+
180
+ - `assigned_gender` (string): Gender assigned to the speaker.
181
+
182
+ ### Data Splits
183
+
184
+ Same splits than in the original [Common Voice corpus](https://commonvoice.mozilla.org/en/datasets)
185
+ The ``reported.tsv`` file hasn't been annotated because it doesn't contain information about the speakers.
186
+
187
+ | split | sentences | manually annotated | % man annot | self-defined accent | % self-d accent | assig accent | % assig accent | self-defined gender | % sd gender | assig gender | % ag gender |
188
+ |---------------------------|------------:|-------------------:|------------:|--------------------:|----------------:|-------------:|---------------:|--------------------:|------------:|-------------:|------------:|
189
+ | annotated_dev.tsv | 16391 | 0 | 0.0 | 2374 | 14.48 | 2391 | 14.59 | 899 | 5.48 | 924 | 5.64 |
190
+ | annotated_train.tsv | 1142932 | 219711 | 19.22 | 666672 | 58.33 | 836843 | 73.22 | 718620 | 62.88 | 854209 | 74.74 |
191
+ | annotated_other.tsv | 489510 | 169583 | 34.64 | 247831 | 50.63 | 423430 | 86.5 | 398199 | 81.35 | 422862 | 86.38 |
192
+ | annotated_test.tsv | 16402 | 0 | 0.0 | 2922 | 17.81 | 3024 | 18.44 | 755 | 4.6 | 785 | 4.79 |
193
+ | annotated_invalidated.tsv | 109232 | 14712 | 13.47 | 65596 | 60.05 | 74399 | 68.11 | 66878 | 61.23 | 75802 | 69.4 |
194
+ | annotated_validated.tsv | 1817954 | 368602 | 20.28 | 1136712 | 62.53 | 1417174 | 77.95 | 1241752 | 68.3 | 1437806 | 79.09 |
195
+
196
+ ## Dataset Creation
197
+
198
+ ### Curation Rationale
199
+
200
+ During 2022, a campaign was launched to promote the Common Voice corpus within the Catalan-speaking community, achieving remarkable success. However, not all participants provided their demographic details such as age, gender, and accent. Additionally, some individuals faced difficulty in self-defining their accent using the standard classifications established by specialists.
201
+
202
+ In order to obtain a balanced corpus with reliable information, we have seen the the necessity of enlisting a group of experts from the University of Barcelona to provide accurate annotations.
203
+
204
+ We hope that this corpus will be a useful contribution to the development of ASR tasks in Catalan.
205
+
206
+ ### Source Data
207
+
208
+ The original data comes from the [Catalan subset of the Common Voice corpus](https://commonvoice.mozilla.org/en/datasets).
209
+
210
+ #### Initial Data Collection and Normalization
211
+
212
+ Starting with [version 13 of the Common Voice corpus](https://commonvoice.mozilla.org/en/datasets) we identified the speakers (273) who have recorded more than 1200 seconds of speech.
213
+ A team of three annotators was tasked with annotating:
214
+ - if all the recordings correspond to the same person.
215
+ - the gender of the speaker
216
+ - the accent of the speaker
217
+ - the quality of the recording
218
+ These annotations have been published in [Zenodo](link).
219
+
220
+ They did a first round annotation, commented their different opinions and then, a second round.
221
+ For this version of the corpus we used the annotations from the second round,
222
+ and selected the IDs (263) where there was agreement between the annotators that correspond to a single person.
223
+
224
+ #### Who are the source language producers?
225
+
226
+ The original data comes from the [Catalan sentences of the Common Voice corpus](https://huggingface.co/datasets/mozilla-foundation/common_voice_16_1/tree/main/transcript/ca).
227
+
228
+ ### Annotations
229
+
230
+ From the original data in the [Common Voice corpus](https://huggingface.co/datasets/mozilla-foundation/common_voice_13_0/tree/main/transcript/ca)
231
+ the following fields have been added:
232
+
233
+ - `annotated_gender`(string): Annotated Gender by the experts team. It can be 'male' or 'female'
234
+
235
+ - `annotated_gender_agreement` (float): Agreement within the annotation team about the gender of the speaker.
236
+
237
+ - `annotated_accent` (string): Annotated Accent by the experts team. The accents considered are: Balearic, Central, Northern, Northwestern, Valencian.
238
+
239
+ - `annotated_accent_agreement` (float): Agreement within the annotation team about the accent of the speaker.
240
+
241
+ - `mean quality` (float): Mean annotated quality of the speakers' recording.
242
+
243
+ - `stdev quality` (float): Deviation in the quality of annotation between annotators.
244
+
245
+ - `propagated_gender` (string): Self-declared gender as indicated in certain recordings by the user.
246
+ If there is a variation in the self-declared gender, "other" is assigned.
247
+
248
+ - `propagated_accents` (string): Self-declared accent as indicated in certain recordings by the user.
249
+ If there is variation between different accents, the majority accent is propagated.
250
+
251
+ - `propagated_accents_normalized` (string): Propagated accent, normalized to the closed-options list used until version 7 (Balearic, Central, Northern, Northwestern, Valencian).
252
+
253
+ - `assigned_accent` (string): Accent assigned to the speaker.
254
+ If the speaker has an expert annotation, it is prioritized.
255
+ Otherwise, the propagated_accents_normalized option will be chosen.
256
+ If expert annotation is not available and ``propagated_accents_normalized`` is not available either, the ``variant`` option is assigned.
257
+ If none of these options are assigned, the field is left empty.
258
+
259
+ - `assigned_gender` (string): Gender assigned to the speaker.
260
+ If the speaker has an expert annotation, it is prioritized.
261
+ Otherwise, the propagated gender is assigned.
262
+ If none of these options are assigned, the field is left empty.
263
+
264
+ #### Annotation process
265
+
266
+ 273 speaker ID have been annotated. However, 10 of them have been discarded because they didn't correspond to a single individual.
267
+
268
+ The selected 263 speakers ID correspond to a total of 361,687 sentences in the validated split (20.18% of the total).
269
+
270
+ The full annotations made by the team of experts are available at [Zenodo](link).
271
+
272
+ #### Who are the annotators?
273
+
274
+ The annotation was entrusted to the [CLiC (Centre de Llenguatge i Computaci贸)](https://clic.ub.edu/en/que-es-clic) team from the University of Barcelona.
275
+ They selected a group of three annotators (two men and one woman), who received a scholarship to do this work.
276
+
277
+ The annotation team was composed of:
278
+
279
+ * 2 male annotators, aged 18-25, L1 Catalan, students in the Catalan Philology degree.
280
+ * 1 female annotator, aged 18-25, L1 Catalan, student in the Modern Languages and Literatures degree, with a focus on Catalan.
281
+ * 1 female supervisor, aged 40-50, L1 Catalan, graduate in Physics and Linguistics, Ph.D. in Signal Theory and Communications.
282
+
283
+ To do the annotation they used a Google Drive spreadsheet
284
+
285
+ ### Personal and Sensitive Information
286
+
287
+ The dataset consists of people who have donated their voice online. You agree to not attempt to determine the identity of speakers in the Common Voice dataset.
288
+
289
+ ## Considerations for Using the Data
290
+
291
+ ### Social Impact of Dataset
292
+
293
+ The dataset consists of people who have donated their voice online.
294
+ You agree to not attempt to determine the identity of speakers in the Common Voice dataset.
295
+
296
+ We hope that this corpus will be a useful contribution to the development of ASR tasks in Catalan
297
+
298
+ ### Discussion of Biases
299
+
300
+ Most of the voices of the common voice in Catalan correspond to men with a central accent between 40 and 60 years old. The aim of this dataset is to provide information that allows to minimize the biases that this could cause.
301
+
302
+ Regarding the content of the recorded sentences, we consider that the Common Voice validation system is efficient in removing those that could produce toxic content.
303
+
304
+ ### Other Known Limitations
305
+
306
+ [N/A]
307
+
308
+ ## Additional Information
309
+
310
+ ### Dataset Curators
311
+
312
+ Language Technologies Unit at the Barcelona Supercomputing Center (langtech@bsc.es)
313
+
314
+ This work has been promoted and financed by the Generalitat de Catalunya through the [Aina project](https://projecteaina.cat/).
315
+
316
+ ### Licensing Information
317
+
318
+ This dataset is licensed under a [CC BY 4.0]((https://creativecommons.org/licenses/by/4.0/)) license.
319
+
320
+ It can be used for any purpose, whether academic or commercial, under the terms of the license.
321
+ Give appropriate credit, provide a link to the license, and indicate if changes were made.
322
+
323
+ ### Citation Information
324
+
325
+ DOI []
326
+
327
+ ### Contributions
328
+
329
+ The annotation was entrusted to the [STeL](https://stel2.ub.edu/el-servei) team from the University of Barcelona.