The Wikispeech projects (Wikispeech 2016, Wikispeech 2019) aim to make Wikipedia and other projects accessible to people that, for different reasons, have difficulties reading. This will be made possible by creating a TTS (text-to-speech) solution that runs on MediaWiki, using FLOSS components. Volunteers will be able to improve the quality of the text-to-speech in true Wikimedia fashion.
Wikipedia contains a lot of information that is freely available to everyone, except that it may not be presented in a way that makes it possible for everyone to take part of it. About 14% of the worlds population is illiterate and about 25-30% learn better from listening than reading. Since most of the information on Wikipedia is written text, this means that it's not available to a significant part of potential readers.
Adding TTS as part of MediaWiki will help these readers to access the information. Having the TTS built into MediaWiki also means that the reader doesn't need any extra equipment or a particularly powerful device to use it as the processing is being handled server-side. While there are a several TTS solutions, the availability of these greatly vary between where you live and what language you speak.
To Further improve Wikispeech Text-to-Speech we will develop tools for collecting speech data, which is required to create new voices. Both tools and any collected data will be freely available to use in other speech technology projects too.
For the MediaWiki extension, see mw:Extension:Wikispeech.
For the Text-to-speech sever, see http://stts-se.github.io/wikispeech/.
For more information on Wikispeech and its design process, see these directory pages:
- Blog post about Dr. Nawar Halabi and the Arabic voice for Wikispeech
- Accessibility/mw:Accessibility for the general efforts in this area
- Extension:PronunciationRecording (for Wiktionary)