# StreamFlow: Streaming Flow Matching with Block-wise Guided Attention Mask for Speech Token Decoding

arXiv: https://arxiv.org/abs/2506.23986 · HTML: https://mlsanity.com/paper/2506.23986/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Audio, speech and music, one of 14 directions in Diffusion models.
- Builds on 21 of its 25 references that are in the map.
- First posted on arXiv 30 Jun 2025.

## Placed in

- Diffusion models / [Audio, speech and music](https://mlsanity.com/diffusion/audio-speech-and-music/)

## Builds on (most cited first, 21 of 21)

- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [Classifier-Free Diffusion Guidance](https://mlsanity.com/paper/2207.12598/) (2022)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [GPT-4o System Card](https://mlsanity.com/paper/2410.21276/) (2024)
- [Scaling Rectified Flow Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2403.03206/) (2024)
- [HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis](https://mlsanity.com/paper/2010.05646/) (2020)
- [FastSpeech 2: Fast and High-Quality End-to-End Text to Speech](https://mlsanity.com/paper/2006.04558/) (2020)
- [Moshi: a speech-text foundation model for real-time dialogue](https://mlsanity.com/paper/2410.00037/) (2024)
- [Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech](https://mlsanity.com/paper/2106.06103/) (2021)
- [UTMOS: UTokyo-SaruLab System for VoiceMOS Challenge 2022](https://mlsanity.com/paper/2204.02152/) (2022)
- [CosyVoice: A Scalable Multilingual Zero-shot Text-to-speech Synthesizer based on Supervised Semantic Tokens](https://mlsanity.com/paper/2407.05407/) (2024)
- [CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language Models](https://mlsanity.com/paper/2412.10117/) (2024)
- [BigVGAN: A Universal Neural Vocoder with Large-Scale Training](https://mlsanity.com/paper/2206.04658/) (2022)
- [GLM-4-Voice: Towards Intelligent and Human-Like End-to-End Spoken Chatbot](https://mlsanity.com/paper/2412.02612/) (2024)
- [Emilia: An Extensive, Multilingual, and Diverse Speech Dataset for Large-Scale Speech Generation](https://mlsanity.com/paper/2407.05361/) (2024)
- [OT-Flow: Fast and Accurate Continuous Normalizing Flows via Optimal Transport](https://mlsanity.com/paper/2006.00104/) (2020)
- [WavChat: A Survey of Spoken Dialogue Models](https://mlsanity.com/paper/2411.13577/) (2024)
- [BASE TTS: Lessons from building a billion-parameter Text-to-Speech model on 100K hours of data](https://mlsanity.com/paper/2402.08093/) (2024)
- [Better speech synthesis through scaling](https://mlsanity.com/paper/2305.07243/) (2023)
- [ECAPA-TDNN Embeddings for Speaker Diarization](https://mlsanity.com/paper/2104.01466/) (2021)
- [The NPU-HWC System for the ISCSLP 2024 Inspirational and Convincing Audio Generation Challenge](https://mlsanity.com/paper/2410.23815/) (2024)

## Built on by (most cited first, 2 of 2)

- [Qwen3-TTS Technical Report](https://mlsanity.com/paper/2601.15621/) (2026)
- [FlashTTS: Fast Streaming TTS with MTP Acceleration and X-pred Mean Flow Distillation](https://mlsanity.com/paper/2606.09141/) (2026)

## Abstract (arXiv)

Recent advancements in discrete token-based speech generation have highlighted the importance of token-to-waveform generation for audio quality, particularly in real-time interactions. Traditional frameworks integrating semantic tokens with flow matching (FM) struggle with streaming capabilities due to their reliance on a global receptive field. Additionally, directly implementing token-by-token streaming speech generation often results in degraded audio quality. To address these challenges, we propose StreamFlow, a novel neural architecture that facilitates streaming flow matching with diffusion transformers (DiT). To mitigate the long-sequence extrapolation issues arising from lengthy historical dependencies, we design a local block-wise receptive field strategy. Specifically, the sequence is first segmented into blocks, and we introduce block-wise attention masks that enable the current block to receive information from the previous or subsequent block. These attention masks are combined hierarchically across different DiT-blocks to regulate the receptive field of DiTs. Both subjective and objective experimental results demonstrate that our approach achieves performance comparable to non-streaming methods while surpassing other streaming methods in terms of speech quality, all the while effectively managing inference time during long-sequence generation. Furthermore, our method achieves a notable first-packet latency of only 180 ms.footnoteSpeech samples: https://dukguo.github.io/StreamFlow/