MeanFlow-TSE: One-Step Generative Target Speaker Extraction with Mean Flow

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Shimizu, Riki, Jiang, Xilin, Mesgarani, Nima
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911329594376192
author Shimizu, Riki
Jiang, Xilin
Mesgarani, Nima
author_facet Shimizu, Riki
Jiang, Xilin
Mesgarani, Nima
contents Target speaker extraction (TSE) aims to isolate a desired speaker's voice from a multi-speaker mixture using auxiliary information such as a reference utterance. Although recent advances in diffusion and flow-matching models have improved TSE performance, these methods typically require multi-step sampling, which limits their practicality in low-latency settings. In this work, we propose MeanFlow-TSE, a one-step generative TSE framework trained with mean-flow objectives, enabling fast and high-quality generation without iterative refinement. Building on the AD-FlowTSE paradigm, our method defines a flow between the background and target source that is governed by the mixing ratio (MR). Experiments on the Libri2Mix corpus show that our approach outperforms existing diffusion- and flow-matching-based TSE models in separation quality and perceptual metrics while requiring only a single inference step. These results demonstrate that mean-flow-guided one-step generation offers an effective and efficient alternative for real-time target speaker extraction. Code is available at https://github.com/rikishimizu/MeanFlow-TSE.
format Preprint
id arxiv_https___arxiv_org_abs_2512_18572
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MeanFlow-TSE: One-Step Generative Target Speaker Extraction with Mean Flow
Shimizu, Riki
Jiang, Xilin
Mesgarani, Nima
Audio and Speech Processing
Target speaker extraction (TSE) aims to isolate a desired speaker's voice from a multi-speaker mixture using auxiliary information such as a reference utterance. Although recent advances in diffusion and flow-matching models have improved TSE performance, these methods typically require multi-step sampling, which limits their practicality in low-latency settings. In this work, we propose MeanFlow-TSE, a one-step generative TSE framework trained with mean-flow objectives, enabling fast and high-quality generation without iterative refinement. Building on the AD-FlowTSE paradigm, our method defines a flow between the background and target source that is governed by the mixing ratio (MR). Experiments on the Libri2Mix corpus show that our approach outperforms existing diffusion- and flow-matching-based TSE models in separation quality and perceptual metrics while requiring only a single inference step. These results demonstrate that mean-flow-guided one-step generation offers an effective and efficient alternative for real-time target speaker extraction. Code is available at https://github.com/rikishimizu/MeanFlow-TSE.
title MeanFlow-TSE: One-Step Generative Target Speaker Extraction with Mean Flow
topic Audio and Speech Processing
url https://arxiv.org/abs/2512.18572