AWhat changed
Developed a streamable quantized audio codec using a multiscale spectrogram adversary, gradient-based loss balancing and optional language-model compression of latent codes.
Scientific scope
Developed a streamable quantized audio codec using a multiscale spectrogram adversary, gradient-based loss balancing and optional language-model compression of latent codes. Builds on prior neural codecs including SoundStream; perceptual reconstruction is lossy. Reported compression and listener preference gains are bounded by evaluated audio, bitrates and compute, not universal fidelity.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Builds on prior neural codecs including SoundStream; perceptual reconstruction is lossy. Reported compression and listener preference gains are bounded by evaluated audio, bitrates and compute, not universal fidelity.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- High Fidelity Neural Audio Compression
- Défossez et al.: EnCodec high-fidelity neural audio compression — author/company project record
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction