Mamba2 Meets Silence: Robust Vocal Source Separation for Sparse Regions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Euiyeon, Choi, Yong-Hoon
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914226330664960
author Kim, Euiyeon
Choi, Yong-Hoon
author_facet Kim, Euiyeon
Choi, Yong-Hoon
contents We introduce a new music source separation model tailored for accurate vocal isolation. Unlike Transformer-based approaches, which often fail to capture intermittently occurring vocals, our model leverages Mamba2, a recent state space model, to better capture long-range temporal dependencies. To handle long input sequences efficiently, we combine a band-splitting strategy with a dual-path architecture. Experiments show that our approach outperforms recent state-of-the-art models, achieving a cSDR of 11.03 dB-the best reported to date-and delivering substantial gains in uSDR. Moreover, the model exhibits stable and consistent performance across varying input lengths and vocal occurrence patterns. These results demonstrate the effectiveness of Mamba-based models for high-resolution audio processing and open up new directions for broader applications in audio research.
format Preprint
id arxiv_https___arxiv_org_abs_2508_14556
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Mamba2 Meets Silence: Robust Vocal Source Separation for Sparse Regions
Kim, Euiyeon
Choi, Yong-Hoon
Sound
Artificial Intelligence
Audio and Speech Processing
We introduce a new music source separation model tailored for accurate vocal isolation. Unlike Transformer-based approaches, which often fail to capture intermittently occurring vocals, our model leverages Mamba2, a recent state space model, to better capture long-range temporal dependencies. To handle long input sequences efficiently, we combine a band-splitting strategy with a dual-path architecture. Experiments show that our approach outperforms recent state-of-the-art models, achieving a cSDR of 11.03 dB-the best reported to date-and delivering substantial gains in uSDR. Moreover, the model exhibits stable and consistent performance across varying input lengths and vocal occurrence patterns. These results demonstrate the effectiveness of Mamba-based models for high-resolution audio processing and open up new directions for broader applications in audio research.
title Mamba2 Meets Silence: Robust Vocal Source Separation for Sparse Regions
topic Sound
Artificial Intelligence
Audio and Speech Processing
url https://arxiv.org/abs/2508.14556