Paper Detail

One-Step Token-to-Waveform Generation with MeanFlow in Latent Space

Zheqi Dai, Guangyan Zhang, Zhen Ye, Jingyu Li, Haolin He, Chunyat Wu, Yiwen Guo, Qiuqiang Kong

arxiv Score 7.3

Published 2026-06-16 · First seen 2026-06-17

General AI

Abstract

Neural audio codecs are central to modern LLM-based Text-to-Speech (TTS) and multimodal systems. As low-bitrate semantic codecs gain prominence, the Token-to-Waveform (Token2Wav) decoder becomes a bottleneck determining both perceptual quality and system efficiency. Conventional multi-step flow-matching decoders offer superior quality but suffer from high inference latency due to iterative sampling, creating a severe quality-speed trade-off. In this paper, we propose a novel Token2Wav architecture that overcomes this limitation by applying MeanFlow in a highly compressed latent space. By modeling the average velocity rather than the instantaneous velocity field, MeanFlow enables true one-step generation. Operating in the latent domain mitigates the memory and stability issues of waveform-level flows, yielding up to a 17$\times$ improvement in Real-Time Factor (RTF) compared to multi-step baselines with negligible quality degradation. Furthermore, we introduce refinement strategies that mitigate latent mismatch, including decoder-only fine-tuning with the MeanFlow generator frozen and end-to-end joint fine-tuning, improving fidelity without increasing inference-time cost. Code and demo are publicly available.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
soon
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@article{dai2026one,
  title = {One-Step Token-to-Waveform Generation with MeanFlow in Latent Space},
  author = {Zheqi Dai and Guangyan Zhang and Zhen Ye and Jingyu Li and Haolin He and Chunyat Wu and Yiwen Guo and Qiuqiang Kong},
  year = {2026},
  abstract = {Neural audio codecs are central to modern LLM-based Text-to-Speech (TTS) and multimodal systems. As low-bitrate semantic codecs gain prominence, the Token-to-Waveform (Token2Wav) decoder becomes a bottleneck determining both perceptual quality and system efficiency. Conventional multi-step flow-matching decoders offer superior quality but suffer from high inference latency due to iterative sampling, creating a severe quality-speed trade-off. In this paper, we propose a novel Token2Wav architectu},
  url = {https://arxiv.org/abs/2606.18072},
  keywords = {eess.AS},
  eprint = {2606.18072},
  archiveprefix = {arXiv},
}

Metadata

{}