Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

TTS behaves oddly in MacOS 15 #48

Closed
mreichhoff opened this issue Oct 26, 2024 · 3 comments · Fixed by #50
Closed

TTS behaves oddly in MacOS 15 #48

mreichhoff opened this issue Oct 26, 2024 · 3 comments · Fixed by #50

Comments

@mreichhoff
Copy link
Owner

The voice is quite fast and robotic, and word highlighting is broken.

The behavior is seen in Chrome, Safari, and Firefox on an (at minimum) an intel Mac.

@mreichhoff
Copy link
Owner Author

mreichhoff commented Nov 19, 2024

not just intel macs!

The voice with name Tingting sounds mostly fine, while the other 8 (!) choices sound awful. It appears simply not setting a voice (instead setting only the utterance's lang and text properties) causes Tingting to be selected on macOS, but will need to verify empty-voice utterances are supported across platforms, or put in OS-specific logic to pick the correct voice

@alfons
Copy link

alfons commented Dec 26, 2024

I've researched a few hours into this and Tingting is the way to go (as of Dec 2024). Reasoning: The only other alternative is to pay for API access for high quality speech synthesis websites, and to either generate the audio on the fly (with async fetch, which most of times proofed to be too slow, especially for flashcards) or to download completely rendered audio and store it with the repository (which is unreasonable given the large amount of examples and combinations.)

@mreichhoff mreichhoff linked a pull request Jan 5, 2025 that will close this issue
@mreichhoff
Copy link
Owner Author

Agreed for now.

I do want to integrate audio files (IIRC Tatoeba has a fair amount, and a bunch of shared anki decks link to voice recordings), plus AI TTS, and possibly the Forvo API at some point, but that's not a today problem.

Hacking in use of Tingting when available in #50. I could go further with OS detection or something, but I'm accepting the risk some other platform has a Tingting voice I wouldn't want.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants