DualCodec: Low‑Frame‑Rate, Semantically‑Enhanced Neural Audio Codec for Speech Generation
DualCodec, a neural audio codec accepted at Interspeech 2025, reduces token frame rates while keeping high audio quality; its code and training scripts are on GitHub. getnews.me/dualcodec-low-frame-rate... #dualcodec #neuralaudio
FlexiCodec Introduces Dynamic Frame-Rate Neural Audio Coding
FlexiCodec varies frame rate from 3 Hz to 12.5 Hz, merging similar frames to cut token count; tests at 6.25 Hz, 8.3 Hz and 12.5 Hz kept semantic content better than baselines. Read more: getnews.me/flexicodec-introduces-dy... #flexicodec #neuralaudio
Watch a full-length workshop on High-Performance Inference for Real-Time Sound Processing from #ADC
See how we implemented Neural Amp Modeler, as well as in-depth info about the GPU Audio SDK, and two audience Q&A sessions. #nvidia #AMD #applesilicon #neuralaudio
www.youtube.com/watch?v=EEKa...
