Time |
Nick |
Message |
00:16 |
|
Noisytoot joined #minetest-dev |
03:24 |
|
fluxionary joined #minetest-dev |
03:43 |
MTDiscord |
<kimapr> there are IRC bridged minetest matrix rooms? |
04:00 |
|
MTDiscord joined #minetest-dev |
05:24 |
|
YuGiOhJCJ joined #minetest-dev |
06:15 |
|
calculon joined #minetest-dev |
08:29 |
|
Evergreen joined #minetest-dev |
11:59 |
MTDiscord |
<josiah_wi> The matrix bridge has been dysfunctional for a while. |
12:25 |
MTDiscord |
<kimapr> they deleted the bridge because it sucked |
14:53 |
|
Noisytoot joined #minetest-dev |
16:14 |
nekobit |
is there an open source text-to-speech library? |
16:14 |
nekobit |
id like to add TTS to Minetest |
16:14 |
MTDiscord |
<bastrabun> There's piper, but that's python as far as I know |
16:23 |
|
nrz joined #minetest-dev |
16:28 |
pgimeno |
there's festival, too |
16:28 |
MTDiscord |
<warr1024> This is what I use. I run it in docker containers, since it's a bit sprawling with its dependencies, but it's way lighter than a lot of the heavy AI stuff I've tried, and actually works. |
16:28 |
MTDiscord |
<warr1024> I found piper to have the best balance between voice quality and actually running and working. |
16:30 |
MTDiscord |
<warr1024> If you're looking for something lightweight and embeddable, I've gotten surprising mileage out of espeak, and it's much lighter weight than piper or other AI-based ones. The voice can be grating if you're picky, but I also found it a lot easier to hear above background noise. |
16:40 |
nekobit |
What if i dont want quality |
16:40 |
nekobit |
i just want to be able to soisoisoisoisoisoisoisoisoi |
16:46 |
MTDiscord |
<kimapr> nekobit: then you use that microsoft voice thing |
16:51 |
nekobit |
that's closed source |
17:07 |
MTDiscord |
<warr1024> if you want to add TTS support directly into the engine, something small and light like whatever espeak does might be your best bet, especially if you have dreams of it someday being considered for merging. |
17:12 |
Noisytoot |
what does espeak use? |
17:16 |
Mantar |
libespeak |
17:17 |
Mantar |
I think pico sounds better, for small TTS libraries |
17:31 |
MTDiscord |
<warr1024> any voice samples from pico out there we can listen to? |
17:33 |
Mantar |
https://p.mort.coffee/pqj.wav |
18:08 |
|
fluxionary joined #minetest-dev |
22:32 |
|
panwolfram joined #minetest-dev |
23:05 |
|
Eragon joined #minetest-dev |
23:53 |
|
behalebabo joined #minetest-dev |
23:53 |
|
jonadab joined #minetest-dev |
23:53 |
|
Pexin joined #minetest-dev |