Quantizer Redundancy and Exact Index Recoding in a Frozen Neural Audio Codec
New in Speech and audio language models; no direction yet. Speech and audio language models
Abstract (arXiv)
Residual scalar quantization can produce highly dependent indices when successive stages reuse identical, unscaled grids. This case study examines that behavior in a frozen neural audio codec and develops an exact recoding procedure that reduces storage while preserving its decoder inputs. Analysis of the repeated-grid recurrence identifies predictable stage sequences; finite-precision checks motivate preserving every original index through explicit exception records. The model weights and reconstruction graph remain fixed. On the studied 62- and 512-file cohorts, a predictor, adaptive coding and explicit file selection reduce complete stored bytes. The final 512-file rate is 129.591 kbps including headers, 25.943% below an original-index XZ control. Original recorded CPU runs establish bitwise latent and waveform preservation; independent readers verify exact index and metadata recovery in subsequent saved-stream experiments. Later coding choices use previously studied cohorts, and fresh encoding requires generating three alternatives. The engineering result is implementation-specific storage reduction with exact index recovery; the underlying lossy reconstruction quality remains unchanged.
Key facts. No citing papers in the map yet (data through 19 Oct 2026). Builds on 10 of its 15 references that are in the map. First posted on arXiv 14 Oct 2026.
Builds on
10 of 15 references.
References · 10
- 2021SoundStream: An End-to-End Neural Audio Codeccited by 988
- 2022High Fidelity Neural Audio Compressioncited by 917
- 2018SDR - half-baked or well done?cited by 626
- 2023High-Fidelity Audio Compression with Improved RVQGANcited by 619
- 2023Finite Scalar Quantization: VQ-VAE Made Simplecited by 498
- 2024Scaling Transformers for Low-Bitrate High-Quality Speech Codingcited by 77
- 2025Robust Residual Finite Scalar Quantization for Neural Compressioncited by 3
- 2026MLVC: Multi-platform Learned Video Codec for Real-World Deploymentcited by 2
- 2026Benchmarking Neural Speech Compression from a Rate-Distortion Perspectivecited by 1
- 2026Opportunistic Conditional Entropy Coding with Frozen Analysis and Synthesis Transformscited by 1
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2025Finite Scalar Quantization Enables Redundant and Transmission-Robust Neural Audio Compression at Low Bit-rates
- 2025Efficient Evaluation of Quantization-Effects in Neural Codecs
- 2026Geometric Iterative Retrieval for Neural Audio Codec Resynthesis
- 2026Switchcodec: Adaptive residual-expert sparse quantization for high-fidelity neural audio coding
- 2026STEPQuant: When and Where Errors Matter in Delta-Rule Recurrent State Quantization
- 2025Two-Dimensional Quantization for Geometry-Aware Audio Coding
- 2025QINCODEC: Neural Audio Compression with Implicit Neural Codebooks
- 2025SACodec: Asymmetric Quantization with Semantic Anchoring for Low-Bitrate High-Fidelity Neural Speech Codecs