nav search
Data Center Software Security Transformation DevOps Business Personal Tech Science Emergent Tech Bootnotes BOFH

Wikipedia to build and give away speech synthesis code

A talking Wikipedia? Yes please, say the visually-impaired

By Richard Chirgwin, 14 Mar 2016

Wikipedia wants users' help to get its articles read out to visually-impaired users.

As many as 125 million Wikipedia users per month either need or prefer to have the text read to them, so Wikipedia Sweden and the KTH Royal Institute of Technology are working together on a crowd-sourced speech platform for the site.

KTH says its software will be published as an open source project, and will plug into any MediaWiki-based site.

While text-to-speech isn't particularly magical, a purpose-built engine would mean the synthesis would work smoothly with the format and structure of articles.

The first language to get the text-to-speech treatment will be Swedish, followed by English, with a proof-of-concept Arabic version.

After those are delivered – around September 2017 – the group will look for users to contribute to getting speech synthesis happening for the remaining 280 languages Wikipedia is published in.

As well as publishing the software, the resulting content will be published under the Wikimedia Commons license.

The Swedish Post and Telecom Authority is financing the project, and speech technology services company STTS is also taking part. KTH is helping the project with work on improving intonation in speech synthesis. ®

The Register - Independent news and views for the tech community. Part of Situation Publishing