It continues to make me sad that #linux has certain areas where it's just... given up.

Speech to text is one of those.

Every other platform has a speech to text engine. Linux has github science projects and demo scripts. Nothing worth using.

Why don't @pipewire developers move onto speech to text? I don't know.

I'm not sure we have the mechanisms in our industry to get programmers to serve users needs like this. It'll only really get solved when it's important to b2b imo.

@doctormo What shortcomings do you see in Speech Dispatcher?
@doctormo @pipewire
I'm sorry, but I don't understand what pipewire has to do with this?
@doctormo @pipewire Yes, in fact the best is mbrola and the others were already better then now with AI, those TTS for Linux can only compete with those for Amiga 500!
@doctormo @pipewire there is espeak, which might not sound good but is super small and works everywhere. I agree with the general sentiment as far as there could be better.
@doctormo @pipewire I have I agree. Not sure the solution, tbh. Hopefully someone will eventually take this up as has happened (for instance) with wireless displays.

@doctormo @pipewire

This seems interesting:
https://github.com/coqui-ai/TTS

Coqui AI is the successor of Mozilla TTS.

GitHub - coqui-ai/TTS: 🐸💬 - a deep learning toolkit for Text-to-Speech, battle-tested in research and production

🐸💬 - a deep learning toolkit for Text-to-Speech, battle-tested in research and production - coqui-ai/TTS

GitHub
@doctormo @pipewire I'm using text-to-speech in GNU Linux, and it suits me! There's espeak as a synthesizer option, for instance, and it works with Orca screen reader, via a system service called Speech Dispatcher. We've had this for years! Speech To Text is another matter, and Android (isn't that Linux Kernel Based), has it. Not sure about the BSD family.

@doctormo @pipewire given up means that somebody actually started to look into it seriously.

The much I like to point out areas where Linux is lacking, phrasing things like this doesn't help at all.

The mechanism you search for is sponsorship. If one think it's needed then either you pay up and sponsor developers to work on it or you implement it yourself if nobody else did so already.

But hoping that programmers "serve" the users is pure entitlement.

@doctormo @pipewire Yeah, that;s true… RHVoice is usable, but that;s the best I can really say
@doctormo @pipewire There;s Mycroft, but it;s not any easy feat to implement their TTS engine for the average user
@doctormo in a debate that I was having with someone else, I made similar point. FOSS in itself doesn't care about users nor their freedom, It only gives you a way to keep the code free , having the code in a repository doesn't mean anything to a non programmer who doesn't have money it is just a bunch of text for them.

FOSS is made by either corporates to fulfil their needs or a volunteer dev fulfilling their needs, very few are made keeping in mind the needs of the users first.
Those who don't know to code, the actual users are at the mercy both of these since they won't have the money to fund such software. And anything they have now is a coincidence that their needs are aligned with the needs of the programmers and the corporates.

Often the response for criticism and requests are met with "patches welcome" kind of replies. I mean if everyone is expected to be a programmer to improve or use FOSS only tools then I am not surprised at the market share of proprietary apps.

One solution might be to form a kind of co-operative or union of users who will collectively fund hire developers and shape the path for the software according to their needs and not what the developers want or like. But this is extremely difficult.