Common Voice Scripted Speech 25.0 - Sindhi

sd — Sindhi (sd)

This datasheet is for cv-corpus-25.0-2026-03-09 of the Mozilla Common Voice Scripted Speech dataset for Sindhi [sd - sd]. The dataset contains 42135 clips representing 47.64 hours of recorded speech (0.36 hours validated) from 31 speakers, recorded from a text corpus of 13,419 sentences.

Language

Sindhi (سنڌي) is an Indo-Aryan language spoken mainly in Sindh, Pakistan, and also in India. It has millions of speakers, a rich literary history, and is written in both Perso-Arabic and Devanagari scripts.

Accents

Demographic information

The dataset includes the following self-declared age and gender distributions. A coverage summary is shown below each table.

Gender

Self-declared gender information. The table shows clip and speaker counts with percentages. Speakers who did not declare a gender are listed as Unspecified. A dash (-) indicates zero.

Gender declared: 5,856 of 42,135 clips (13.9%), 1 of 31 speakers (3.2%)

Age

Self-declared age information. The table shows clip and speaker counts with percentages. Speakers who did not declare an age are listed as Unspecified. A dash (-) indicates zero.

Age declared: 41,442 of 42,135 clips (98.4%), 4 of 31 speakers (12.9%)

Data splits for modelling

Clip buckets

Training splits

Training split coverage: 321 of 321 validated clips (100.0%)

The dataset contains 321 validated, 232 invalidated, and 41582 unresolved clips. The average clip duration is 4.07 seconds.

Text corpus

The Sindhi corpus consists of collected texts from newspapers, and social media. It contains more then one lacs sentences. The texts cover different domains, including literature, news, education, and everyday communication.

Validated sentences: 13,348

The corpus contains 13,419 sentences: 13,348 validated and 71 unvalidated (69 pending review, 2 rejected), with 9 reported for review.

Writing system

The Sindhi corpus is written in the Perso-Arabic script, which is the standard script used by Sindhi newspapers in Pakistan. It includes additional letters that represent Sindhi-specific sounds not found in standard Arabic.

Symbol table

The Sindhi Perso-Arabic script (used in newspapers) has 52 letters. Here is the full alphabet list separated by spaces: ا ب ٻ ڀ پ ت ٿ ٽ ٺ ث ج جه ڄ چ ڇ ح خ د ڌ ڏ ڊ ڍ ذ ر ڙ ڍ ز س ش ص ض ط ظ ع غ ف ڦ ق ڪ گ ڳ گه ڱ ل ڻ ن ڃ م و ء ه ة ي ئ

Sample

There follows a randomly selected sample of five sentences from the corpus.

1. اقامت جو مطلب آهي

2. اهو هڪ ڌماڪو سڏيو ويندو آهي.

3. فنا جي ذائقي سان فرصت جي مٽي، اي انتظار

4. هڪ ڀيرو ٻيهر ڪرونا ڪيس تيزي سان سامهون اچڻ شروع ٿي ويا

5. جڏهن ته فعل جي ڪا به معنيٰ ناهي.

Sources

Text domains

General, Media and Entertainment, News and Current Affairs

Processing

The corpus was created by collecting texts from different Sindhi newspapers. The articles were gathered, cleaned to remove formatting issues, and then organized into a structured dataset for analysis.

Recommended post-processing

It is recommended to clean the text for duplicate articles, normalize spellings, and remove unwanted symbols or formatting. Tokenization and sentence segmentation may also be useful for better analysis.

Fields

Clips

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

• client_id - hashed UUID of a given user

• path - relative path of the audio file

• text - supposed transcription of the audio

• up_votes - number of people who said audio matches the text

• down_votes - number of people who said audio does not match text

• age - age of the speaker1

• gender - gender of the speaker1

• accents - accents of the speaker1

• variant - variant of the language1

• segment - if sentence belongs to a custom dataset segment, it will be listed here

• prompt_upvotes - number of upvotes the sentence prompt received

• prompt_reports - number of reports the sentence prompt received

• is_edited - whether the clip's transcription has been edited

validated_sentences.tsv

The validated_sentences.tsv file contains one row per validated sentence in the text corpus:

• sentence_id - unique identifier for the sentence

• sentence - the sentence text

• variant - the variant of the language

• sentence_domain - the domain(s) the sentence belongs to

• source - the source the sentence was collected from

• is_used - whether the sentence is still in circulation for recording

• clips_count - number of clips recorded for this sentence

unvalidated_sentences.tsv

The unvalidated_sentences.tsv file contains one row per unvalidated sentence in the text corpus:

• sentence_id - unique identifier for the sentence

• sentence - the sentence text

• variant - the variant of the language

• sentence_domain - the domain(s) the sentence belongs to

• source - the source the sentence was collected from

• up_votes - number of upvotes the sentence received

• down_votes - number of downvotes the sentence received

• status - current status of the sentence (pending or rejected)

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

• Speak

• Write

• Listen

• Review

Acknowledgements

Datasheet authors

Common Voice Community

Funding

This dataset was created with funding support from Mozilla. Special acknowledgments to Meesam Alam (meesum.alam12@gmail.com) for contributions and support.

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 ↩3 ↩4