# A High-Quality and Low-Complexity Streamable Neural Speech Codec with Knowledge Distillation

arXiv: https://arxiv.org/abs/2509.13670 · HTML: https://mlsanity.com/paper/2509.13670/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Speech and audio distillation, one of 12 directions in Knowledge distillation.
- Builds on 19 of its 25 references that are in the map.
- First posted on arXiv 17 Sep 2025.

## Placed in

- Knowledge distillation / [Speech and audio distillation](https://mlsanity.com/distill/speech-and-audio-distillation/)

## Builds on (most cited first, 19 of 19)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [SoundStream: An End-to-End Neural Audio Codec](https://mlsanity.com/paper/2107.03312/) (2021)
- [Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers](https://mlsanity.com/paper/2301.02111/) (2023)
- [ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders](https://mlsanity.com/paper/2301.00808/) (2023)
- [High Fidelity Neural Audio Compression](https://mlsanity.com/paper/2210.13438/) (2022)
- [LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speech](https://mlsanity.com/paper/1904.02882/) (2019)
- [AudioLM: a Language Modeling Approach to Audio Generation](https://mlsanity.com/paper/2209.03143/) (2022)
- [Finite Scalar Quantization: VQ-VAE Made Simple](https://mlsanity.com/paper/2309.15505/) (2023)
- [High-Fidelity Audio Compression with Improved RVQGAN](https://mlsanity.com/paper/2306.06546/) (2023)
- [SpeechTokenizer: Unified Speech Tokenizer for Speech Large Language Models](https://mlsanity.com/paper/2308.16692/) (2023)
- [ViSQOL v3: An Open Source Production Ready Objective Speech and Audio Metric](https://mlsanity.com/paper/2004.09584/) (2020)
- [High-Quality, Low-Delay Music Coding in the Opus Codec](https://mlsanity.com/paper/1602.04845/) (2016)
- [APCodec: A Neural Audio Codec with Parallel Amplitude and Phase Spectrum Encoding and Decoding](https://mlsanity.com/paper/2402.10533/) (2024)
- [Low-latency Speech Enhancement via Speech Token Generation](https://mlsanity.com/paper/2310.08981/) (2023)
- [APCodec+: A Spectrum-Coding-Based High-Fidelity and High-Compression-Rate Neural Audio Codec with Staged Training Paradigm](https://mlsanity.com/paper/2410.22807/) (2024)
- [MDCTCodec: A Lightweight MDCT-based Neural Audio Codec towards High Sampling Rate and Low Bitrate Scenarios](https://mlsanity.com/paper/2411.00464/) (2024)
- [ERVQ: Enhanced Residual Vector Quantization with Intra-and-Inter-Codebook Optimization for Neural Audio Codecs](https://mlsanity.com/paper/2410.12359/) (2024)
- [A Streamable Neural Audio Codec with Residual Scalar-Vector Quantization for Real-Time Communication](https://mlsanity.com/paper/2504.06561/) (2025)

## Abstract (arXiv)

While many current neural speech codecs achieve impressive reconstructed speech quality, they often neglect latency and complexity considerations, limiting their practical deployment in downstream tasks such as real-time speech communication and efficient speech compression. In our previous work, we proposed StreamCodec, which enables streamable speech coding by leveraging model causalization and a scalar-vector-combined quantization strategy, but its reconstructed quality and complexity still have room for improvement. Therefore, this paper proposes an improved iteration of StreamCodec, named StreamCodec2. The StreamCodec2 supports streamable and lightweight speech coding by adopting a fully causal architecture and reducing the convolutional channels. To compensate for the speech quality degradation caused by model causalization and pruning, we introduce a non-causal, high-complexity teacher codec to guide the training of StreamCodec2 through knowledge distillation. Experimental results demonstrate that our proposed StreamCodec2, trained with the knowledge distillation strategy, can achieve high-quality speech reconstruction while maintaining low latency (only 20 ms), low computational complexity (only 910 MFLOPs), and low model complexity (only 5.4 M parameters).