Is there any open-source Text to Speech (TTS) option available in Julia that:
can save to (.ogg) file, and
has a licence permitting commercial use
Doesn’t look like Windows and MacOS allow free commercial use.
Looks like there are open source options like Mozilla TTS available, but I don’t have the skills to turn it into a package, which would be the ideal solution.
Best might be nr. 2 there, GPL3-licenced (nr. 1 is Java-based):
When selecting the best free text-to-speech software is best for you depends on a range of factors (not to mention personal preference). … We also want to test the accessibility features of these tools to see how they work for every kind of user out there. We have highlighted, for instance, whether certain software offer dyslexic-friendly fonts, such as the number two on our list, Natural Reader.
AI/deep learning based are going to be best like this one (but you could also argue a, smaller, package that accesses platform-provided TTS, that will likely improve over time, but will not be consistent across platforms):
OCTAVE TTS, the first text-to-speech system built on LLM intelligence. Unlike conventional TTS […]
Hume’s state-of-the-art expression measurement models for the voice, face, and language are built on 10+ years of research and advances in semantic space theory pioneered by Alan Cowen.
I confirmed that last one even supports Icelandic (but it’s not perfect, as expected, maybe for most other languages).
TTS is available in platforms like Windows and Android, but will be inferior for a while (also likely free software), but going forward it will be a perfect standard feature to be expected, like font rendering now, so then for sure relying on it will be better than bundling TTS in a package, or calling a web API.
There’s no need to do it in Julia from scratch, probably worse (reuse good stuff out there):
I couldn’t confirm TTS (yet) available for Julia (easily, without Python involvment, though pretty easy with PythonCall.jl), but doing the reverse problem is already available, without involving Python:
[Whisper was state-of-the-art, then updated, but is no longer; new SOTA as of last week or so (from China if I recall), this is till a very active research area, more so than TTS, but TTS is also moving along.]
A lot is available already:
I did find some false positives while looking into this: