search

natlamir / PiperUI

A UI for the Piper TTS
42 stars 7 forks source link

longform audio #10

Open mapleroyal opened 6 months ago

mapleroyal commented 6 months ago

Would you consider making a feature where it can generate book-length audio instead of short sentences?

I know everyone just wants to use this to voice their AI waifus, but it may be the solution I've been looking for for a LONG time to convert ebooks to audiobooks.

As a bonus, I don't know if you're aware but there's a pretty big industry for ebook TTS apps on mobile phones. Someone could put all those apps out of business if they just turned this into a simple mobile app and included a couple high quality voices.

yesuda commented 1 month ago

I've been working on my own version of a GUI for piper and I'll have multi-line supported, but for something like an audiobook, wouldn't you prefer to use a bigger cloud tts like openais? https://github.com/p0n1/epub_to_audiobook

mapleroyal commented 1 month ago

I've been working on my own version of a GUI for piper and I'll have multi-line supported, but for something like an audiobook, wouldn't you prefer to use a bigger cloud tts like openais? https://github.com/p0n1/epub_to_audiobook

Oh ya 100% if it was free and local and I could initiate it and go off to do something else and return and it would have a completed output audio file of the full text. But I think any major corporate TTS service is none of those.

yesuda commented 1 month ago

The repo I linked should check off 2 of those 3 things, but I'll try adding epub pdf support to my own gui

mapleroyal commented 1 month ago

I may be misunderstanding you? The repo you linked appears to use non-free, non-local models…?

yesuda commented 1 month ago

i haven't actually used that program but im pretty sure edge-tts doesn't need an api key or anything, so it's not local but free and turns a full text into an audio file