mastodon - Link to source
Look y'all, since a few of you have asked about NVDA 2026.1 and my speech synths, I'm not going to start work on this until at least betas are out. @fastfinge has done a phenominal work starting around this, and his code helps me understand how such a thing would be done, but for one, I don't want an inter-process thing to be written in C++ but rather Python, sort of how he has it. Until #NVDASR develops their own adapter that developers can reuse for any DLL and just adapt function names or signatures to the workings of that speech synth, that process would become much, much easier. If that host process defined an API shape existing speech synths just hook into, I'm all for it. Then I'll begin that work, but not until then. Thanks for getting it, or not.
#nvdasr @🇨🇦Samuel Proulx🇨🇦
This entry was edited (1 month ago)
in reply to Luis Carlos

mastodon - Link to source

Tamas G

@luiscarlosgonzalez LOL between Brailab, 3 versions of Eloquence, SoftVoice, SMPSoft, Flexvoice, I just will have my hands busy redoing all of it. Not going to look forward to that work, if there's a reason I get GPT 5 pro again, that'll be it because by hand to recode so many synths... Hopefully those other devs come out to do the work too, but it's never a 100% guarantee and NVDA can't commit devs to that work nor are they obligated to continue it after the initial release if they cannot. That's just the nature of open-source stuff, I myself won't have time to pick it up but I know @mush42 has been looking for new maintainers for his work. @fastfinge @ZBennoui
@Musharraf @Luis Carlos @Zach Bennoui @🇨🇦Samuel Proulx🇨🇦
in reply to Luis Carlos

iceshrimp.net - Link to source

🇨🇦Samuel Proulx🇨🇦

If you want to do that with gemini-cli, make sure to give it access to a directory with the NVDA Addon developer guide, the API documentation for the synth you’re coding, and tell it where the NVDA repository is. The prompt should include each file name, and what it is. Then Gemini can read the documentation without searching the web, and that will save you a lot of time and tokens. If you have any header files or other bindings, also give it those.
in reply to Luis Carlos

iceshrimp.net - Link to source

🇨🇦Samuel Proulx🇨🇦

Right but Sherpa onnx might replace that. Or the various Microsoft AI providers. We don’t need to care about cross platform for NVDA. And so few people in this community have the expertise to write our own inference framework, I feel strongly that we really need to use something off the shelf. Otherwise, history will repeat itself, and the person working on the framework will get a job or otherwise be forced to drop the project. I can only think of four blind people who even know rust. And I’m not one of them.And the others all have full time jobs.
in reply to Luis Carlos

iceshrimp.net - Link to source

🇨🇦Samuel Proulx🇨🇦

Exactly that has already happened, and I was forced to pay money to strip Camlorn’s custom 3d audio library out of unspoken. Because he’s not maintaining that anymore, and he’s the only one in our community with that particular set of skills. If we as blind people want to write software for ourselves that will last, we have to stop depending on custom low level frameworks and make the things that already exist work for us.
⇧