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.

Sign up for GitHub

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

Jump to bottom

Question about bitrate choices #37

Open
VivekPanyam opened this issue Mar 3, 2023 · 3 comments
Open

Question about bitrate choices #37

VivekPanyam opened this issue Mar 3, 2023 · 3 comments
Labels
question Further information is requested

Comments

@VivekPanyam
Copy link

❓ Questions

Hello! Do you intend to train models with a target bitrate of above 24kbps? I didn't see anything in the paper, but maybe I missed it.

I'd be curious to see how 48 and 96kbps models compare to mp3s at higher bitrates.

Thanks for the great work!
@VivekPanyam VivekPanyam added the question Further information is requested label Mar 3, 2023
@adefossez
Copy link
Contributor

this kind of architecture is really good at low bitrates but fail to be competitive at higher bitrates. specifically, it will not reach a state where the audio is indistinguishable from the uncompressed audio as you increase the bandwidth, unlike mp3.

@VivekPanyam
Copy link
Author

That makes sense, thanks! Are you aware of any papers or architectures that can operate at higher quality levels while still having some sort of discrete representation internally (via RVQ or something else)?

@listener17
Copy link

Have you found any such papers? For audio/speech/image/video?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
question Further information is requested
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@VivekPanyam @adefossez @listener17