Login
Login

Common Voice

{{Short description|Voice dataset by Mozilla}}

{{Infobox software

| name = Common Voice

| logo = Common Voice Banner2.png

| website = [https://commonvoice.mozilla.org/ commonvoice.mozilla.org]

| caption = Teach machines how real people speak

| language = Multilingual ([https://voice.mozilla.org/languages List of languages])

| developer = Mozilla Foundation

| released = {{Start date and age|2017|6|19}}

| repo = {{URL|https://github.com/common-voice/common-voice}}

| license = Creative Commons CC0

}}

Common Voice is a crowdsourcing project started by Mozilla to create a free database for speech recognition software. The project is supported by volunteers who record sample sentences with a microphone and review recordings of other users. The transcribed sentences are collected in a voice database available under the public domain license CC0.{{Cite web |title=Mozilla Common Voice |url=https://commonvoice.mozilla.org/en/datasets |access-date=2024-10-06 |website=commonvoice.mozilla.org |language=en}} This license ensures that developers can use the database for voice-to-text applications without restrictions or costs.

Aims

Common Voice aims to provide diverse voice samples. According to Mozilla's Katharina Borchert, many existing projects took datasets from public radio or otherwise had datasets that underrepresented both women and people with pronounced accents.{{cite news |date=11 January 2020 |title=Why do we gender AI? Voice tech firms move to be more inclusive |work=The Guardian |url=https://www.theguardian.com/technology/2020/jan/11/why-do-we-gender-ai-voice-tech-firms-move-to-be-more-inclusive |accessdate=19 April 2020 |archive-date=19 December 2022 |archive-url=https://web.archive.org/web/20221219033558/https://www.theguardian.com/technology/2020/jan/11/why-do-we-gender-ai-voice-tech-firms-move-to-be-more-inclusive |url-status=live }}

History

{{update section|date=October 2024}}

At the beginning of 2022, Bengali.AI partnered with Common Voice to launch "Bangla Speech Recognition" project that aims to make machines understand Bangla language. 2000 hours of voice was collected with aim for higher than 10,000 hours.{{Cite web |date=2022-12-23 |title=Bengali.AI: Democratising AI research in Bangla |url=https://www.tbsnews.net/features/panorama/bengaliai-democratising-ai-research-bangla-556458 |access-date=2022-12-25 |website=The Business Standard |language=en |archive-date=2022-12-24 |archive-url=https://web.archive.org/web/20221224094036/https://www.tbsnews.net/features/panorama/bengaliai-democratising-ai-research-bangla-556458 |url-status=live }}

Voice database

The first dataset was released in November 2017. More than 20,000 users worldwide had recorded 500 hours of English sentences.{{cite web|url=https://blog.mozilla.org/blog/2017/11/29/announcing-the-initial-release-of-mozillas-open-source-speech-recognition-model-and-voice-dataset|title=Announcing the Initial Release of Mozilla’s Open Source Speech Recognition Model and Voice Dataset|date=November 29, 2017|website=blog mozilla.org|access-date=November 19, 2019|archive-date=November 29, 2017|archive-url=https://web.archive.org/web/20171129164616/https://blog.mozilla.org/blog/2017/11/29/announcing-the-initial-release-of-mozillas-open-source-speech-recognition-model-and-voice-dataset|url-status=live}}

In February 2019, the first batch of languages was released for use. This included 18 languages: English, French, German and Mandarin Chinese, but also less prevalent languages as Welsh and Kabyle. In total, this included almost 1,400 hours of recorded voice data from more than 42,000 contributors.{{cite web|url=https://venturebeat.com/2019/02/28/mozilla-updates-common-voice-dataset-with-1400-hours-of-speech-across-19-languages|title=Mozilla updates Common Voice dataset with 1,400 hours of speech across 18 languages|website=VentureBeat|date=February 28, 2019|access-date=November 19, 2019|archive-date=March 4, 2019|archive-url=https://web.archive.org/web/20190304213117/https://venturebeat.com/2019/02/28/mozilla-updates-common-voice-dataset-with-1400-hours-of-speech-across-19-languages|url-status=live}}

As of July 2020 the database has amassed 7,226 hours of voice recordings in 54 languages, 5,591 hours of which has been verified by volunteers.{{cite web |title=Mozilla Common Voice updates will help train the ‘Hey Firefox’ wakeword for voice-based web browsing |url=https://venturebeat.com/2020/07/01/mozilla-common-voice-updates-will-help-train-the-hey-firefox-wakeword-for-voice-based-web-browsing/ |website=VentureBeat |access-date=1 April 2021 |archive-url=https://web.archive.org/web/20210310211235/https://venturebeat.com/2020/07/01/mozilla-common-voice-updates-will-help-train-the-hey-firefox-wakeword-for-voice-based-web-browsing/ |archive-date=March 10, 2021 |date=1 July 2020}}

In May 2021, following the work to add Kinyarwanda, they received a grant to add Kiswahili.{{Cite web|date=2021-05-25|title=Mozilla Common Voice Receives $3.4 Million Investment to Democratize and Diversify Voice Tech in East Africa|url=https://foundation.mozilla.org/en/blog/mozilla-common-voice-receives-34-million-investment-to-democratize-and-diversify-voice-tech-in-east-africa/|access-date=2021-06-03|website=Mozilla Foundation|language=en|archive-date=2022-12-19|archive-url=https://web.archive.org/web/20221219033600/https://foundation.mozilla.org/en/blog/mozilla-common-voice-receives-34-million-investment-to-democratize-and-diversify-voice-tech-in-east-africa/|url-status=live}}

In September 2022, it was announced that the Twi language of Ghana was the 100th language to be added to the Mozilla Common Voice database.{{cite web |last1=Onukwue |first1=Alexander |title=Ghana’s most popular language is now on Mozilla Common Voice |url=https://qz.com/ghana-s-most-popular-language-will-be-available-to-more-1849572359 |website=Quartz |access-date=3 October 2022 |language=en-us |date=23 September 2022 |archive-date=2 December 2022 |archive-url=https://web.archive.org/web/20221202202435/https://qz.com/ghana-s-most-popular-language-will-be-available-to-more-1849572359 |url-status=live }}

{{As of|October 2022}}, Mozilla Common Voice officially collects voice data for the following languages:{{cite web |title=Languages |url=https://commonvoice.mozilla.org/en/languages |website=commonvoice.mozilla.org |access-date=4 October 2022 |language=en |archive-date=24 December 2022 |archive-url=https://web.archive.org/web/20221224203252/https://commonvoice.mozilla.org/en/languages |url-status=live }}

{{Div col|colwidth=15em}}

{{Div col end}}

See also

References

{{Mozilla}}

Category:Speech recognition software

Category:Datasets in machine learning

Category:Internet properties established in 2017