Thanks. I know that it's not the same (though I believe that Common Voice builds databases that can be used for many goals ?) but basically if we have a linguist for Common Voice we can ask him to help for speech synthesis too (he was highly motivated with helping as he is passionate in promoting the use of local languages in modern usages; also he's keenly aware of how getting speech interfaces in local languages, and not only text interfaces, will completely revolutionnarize the way illiterate populations can use the now ubiquitous IT tools). I've begun parsing the documentation of espeak-ng but that's a lot of information to process, and the meeting with ABPAM is supposed to be tomorrow. Could you please just explain what are the lower-level contributions that would help adding a new language ? The links-in-the-link that you provided say "In many cases it should be fairly easy to add a rough implementation of a new language, hopefully enough to be intelligible. After that it's a gradual process of improvement." which seems to me to be fine for what we're trying to do (give the kids a proof of concept to keep them motivated, then teach them how to improve it with time). But from what I glanced, I see only work requiring either computer knowledge or linguistic skills. As I wrote earlier, we have people in the team that can provide that knowledge and probably teach how to do it to the more advanced kids, but it would be very important to have also "grunt work" to give to all kids as an entry-level task, so as to help them understand that contribution is possible to everybody, and to keep a team spirit. I wasn't able to identify that sort of "grunt work" from the quick look I had at the documentation, though. And I didn't understand whether it would be required to have people pronouncing phrases in Mooré or Fulfuldé or whatever (like Common Voice requests it) or whether the computer would only require phonemes (nor where these phonemes would come from) ? ------- Original Message ------- On Sunday, March 6th, 2022 at 4:13 PM, Samuel Thibault <sthibault@debian.org> wrote: > Hello, > > LibreFaso, le dim. 06 mars 2022 14:48:00 +0000, a ecrit: > > > I had a meeting with a local linguist who agreed to help us develop a Mooré (and probably Bamanaan an Fulfulde) version of Common Voice, so actually helping the kids develop a local language version of speech synthesis is definitely a possibility. > > AFAIK Common Voice provides speech recognition, not speech synthesis. > > For speech synthesis the most common project would be espeak-ng > > https://github.com/espeak-ng/espeak-ng/ > > They have a contribution guide: > > https://github.com/espeak-ng/espeak-ng/blob/master/docs/contributing.md > > Samuel
Attachment:
signature.asc
Description: OpenPGP digital signature