Common Voice Spontaneous Speech 3.0 - Catalan

català — Catalan (ca)

This datasheet is for sps-corpus-3.0-2026-03-09 of the Mozilla Common Voice Spontaneous Speech dataset for Catalan [català - ca]. The dataset contains 149 clips representing 0.65 hours of recorded speech (0.49 hours validated) from 15 speakers.

Language

Catalan is a Romance language spoken by about 9 milion people mainly on the Mediterranean coast of the Iberian Peninsula.

It is an official language, along with Spanish or Castilian, in Catalonia, the Balearic Islands and the Valencian Community (where it is also called Valencian), while it is the only official language of the Principality of Andorra. It is also spoken, and has some administrative recognition, without reaching official status, in the eastern part of the autonomous community of Aragon, in the French department Pyrénées-Orientales (Eastern Pyrenees) and in the city of Alghero, on the island of Sardinia (Italy).

The language evolved from Vulgar Latin in the Middle Ages.

Variants

The main variants of Catalan are:

• Central [ca-central]: It is the variant with the most speakers, as it encompasses the metropolitan area of Barcelona, extending to the region of Girona and the eastern half of Tarragona

• Balearic [ca-balear]: The variant used in the Balearic Islands

• Nord-Occidental [ca-nwestern]: Spoken in Andorra, Lleida and the western half of Tarragona in Catalonia, and the eastern part of Aragon

• Septentrional [ca-northern]: Corresponds to the area of Roussillon and the northern part of Girona

• Valencian: Spoken in the Valencian comunity, where it's also known as "Valencian"

  • Valencià meridional [ca-valencia-southern]

  • Alacantí [ca-valencia-alacant]

  • Valencià septentrional [ca-valencia-northern]

  • Tortosí [ca-valencia-tortosi]

  • Valencià central [ca-valencia-central]

• Alguerese [ca-algueres]: Spoken in the city of Alghero, in Sardinia

Data splits for modelling

The dataset clips are categorised by transcription status and training-set assignment. The following tables summarise the distribution.

Audio clips

Training splits

Training split coverage: 0 of 123 transcribed & validated clips (0.0%)

Transcriptions

Transcription status

Writing system

Catalan is written using the Latin alphabet (abcdefghijklmnopqrstuvwxyz), with the special characters ç and l·l. In addition, vowels can be accented (à, è, é, í, ò, ó, ú, ü, ï). The characters - (hyphen) and ' (apostrophe) are also part of Catalan orthography.

Symbol table

a b c ç d e f g h i j k l m n o p q r s t u v w x y z à è é í ò ó ú ï ü

Samples

Questions

There follows a randomly selected sample of questions used in the corpus.

1. Quina tasca de voluntariat tʼagradaria provar i per què?

2. Què significa lʼèxit per a tu?

3. Quina és la teva festa preferida i per què?

4. Quins elements fan que un petit comerç local et resulti especialment atractiu?

5. Com valores el fet de fer la migdiada els caps de setmana?

Responses

There follows a randomly selected sample of transcribed responses from the corpus.

1. Mira, em el trobo d'un, diguem, el que a mi m'emociona dels castells, perquè en allà veig alguna cosa preciosa que fem entre tots, és allà un un individu allà és rellevant, la importància, el castell, esdevé quan tots nem...

2. Les ganes de fer vacances

3. De tot. M'encanta aprendre. M'agradaria molt aprendre més coses i tenir temps per aprendre més coses. Ja tinc una certa edat i, ara estic aprenent italià, que és una llengua que m'agrada molt i, penso que, quan em jubili, una part del temps que guanyaré el dedicaré a aprendre encara més coses. Com que soc més de lletres, doncs, segurament seran, jo que sé, noves llengües o, m'apuntaré a cursos o a seminaris d'història que m'ha és el que més m'ha agradat a la vida i és el que he estudiat o eh, no ho sé, coses així: música... Però, segur, segur que dedicaré molt de temps a seguir aprenent. Penso que és essencial durant la vida aprendre fins al final.

4. Diria que l'amabilitat i el tracte dels comerciants, també el coneixement mutu que es pot tindre quan és un xicotet comerç entre el comercial i el client i... pense que també és important la varietat i tindre preus raonables.

5. Molta gent comprant al voltant de les botigues, parlant entre ells.

Recommended post-processing

It is recommended to normalize instances of the geminate L, which can take the equivalent forms of l·l or ŀl.

Fields

Each row of a tsv file represents a single audio clip, and contains the following information:

• client_id - hashed UUID of a given user

• audio_id - numeric id for audio file

• audio_file - audio file name

• duration_ms - duration of audio in milliseconds

• prompt_id - numeric id for prompt

• prompt - question for user

• transcription - transcription of the audio response

• votes - number of people that who approved a given transcript

• age - age of the speaker1

• gender - gender of the speaker1

• language - language name

• split - for data modelling, which subset of the data does this clip pertain to

• char_per_sec - how many characters of transcription per second of audio

• quality_tags - some automated assessment of the transcription--audio pair, separated by |

  • transcription-length - character per second under 3 characters per second

  • speech-rate - characters per second over 30 characters per second

  • short-audio - audio length under 2 seconds

  • long-audio - audio length over 5 minutes

Get involved

Community links

• Common Voice translators on Pontoon

• Common Voice Communities

Discussions

• Common Voice on Matrix

• Common Voice on Discourse

• Common Voice on Discord

• Common Voice on Telegram

Contribute

• Contribute questions

• Validate questions

• Answer questions

• Transcribe recordings

• Validate transcriptions

Acknowledgements

Datasheet authors

• Carme Armentano <carme.armentano@bsc.es>

Licence

This dataset is released under the Creative Commons Zero (CC-0) licence. By downloading this data you agree to not determine the identity of speakers in the dataset.

Footnotes

1. For a full list of age, gender, and accent options, see the demographics spec. These will only be reported if the speaker opted in to provide that information. ↩ ↩2