Aria-K-Alethia / BigCodec

Official implementation of the paper "BigCodec: Pushing the Limits of Low-Bitrate Neural Speech Codec"
https://aria-k-alethia.github.io/bigcodec-demo/
MIT License
84 stars 4 forks source link

multiple codebooks? #7

Open wincing2 opened 1 month ago

wincing2 commented 1 month ago

Current BigCodec configuration is of 80Hz, which is too high for my TTS task in terms of latency. Is it possible to change the configuration to 2 codebooks and reduce token rate in half?

I've tried low token rate with single codebook. The reconstrction CER is much higher than that of 80hz. So I want to increase the number of codebooks.

Aria-K-Alethia commented 1 month ago

Hi,

I think it's easy and straightforward to implement a multiple-codebook BigCodec based on my released code.