# SAC: Neural Speech Codec with Semantic-Acoustic Dual-Stream Quantization

arXiv: https://arxiv.org/abs/2510.16841 · HTML: https://mlsanity.com/paper/2510.16841/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 19 Oct 2025.

## Built on by (most cited first, 8 of 8)

- [dots.tts Technical Report](https://mlsanity.com/paper/2606.07080/) (2026)
- [Habibi: Laying the Open-Source Foundation of Unified-Dialectal Arabic Speech Synthesis](https://mlsanity.com/paper/2601.13802/) (2026)
- [X-VC: Zero-shot Streaming Voice Conversion in Codec Space](https://mlsanity.com/paper/2604.12456/) (2026)
- [DSA-Tokenizer: Disentangled Semantic-Acoustic Tokenization via Flow Matching-based Hierarchical Fusion](https://mlsanity.com/paper/2601.09239/) (2026)
- [FlashTTS: Fast Streaming TTS with MTP Acceleration and X-pred Mean Flow Distillation](https://mlsanity.com/paper/2606.09141/) (2026)
- [Content is What Remains: Invariant Speech Tokenization from Parallel Utterances](https://mlsanity.com/paper/2607.19033/) (2026)
- [Stable Autoregressive Speech Generation with Low-Frame-Rate High-Dimensional Continuous Tokens](https://mlsanity.com/paper/2607.29363/) (2026)
- [TokenMapper: A Step Toward Interoperable Speech Token Translation](https://mlsanity.com/paper/2609.12563/) (2026)

## Abstract (arXiv)

Speech codecs that convert continuous speech signals into discrete tokens have become essential for speech language models. However, existing codecs struggle to balance high-quality reconstruction with semantically rich representations, limiting their effectiveness in both generative and understanding tasks. In this work, we propose SAC, a neural speech codec with semantic-acoustic dual-stream quantization. By disentangling semantic and acoustic modeling into two dedicated streams, SAC enables each to be optimized for its respective role. Comprehensive evaluations show that SAC achieves strong reconstruction performance across diverse bitrates under both clean and noisy conditions, with particularly high scores on UTMOS and WER, indicating superior naturalness and intelligibility. Moreover, SAC substantially surpasses prior codecs in semantic representation, approaching the level of continuous self-supervised embeddings. When used as a tokenizer for LLM-based text-to-speech, SAC enables a single-stage autoregressive (AR) TTS model that clearly outperforms state-of-the-art AR systems. Our disentanglement analysis further validates the effectiveness of the dual-stream design, offering new potential for controllable speech generation.