Guardado en:
Detalles Bibliográficos
Autores principales: Du, Hui-Peng, Ai, Yang, Jiang, Xiao-Hang, Zheng, Rui-Chen, Ling, Zhen-Hua
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:https://arxiv.org/abs/2601.12757
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911384227282944
author Du, Hui-Peng
Ai, Yang
Jiang, Xiao-Hang
Zheng, Rui-Chen
Ling, Zhen-Hua
author_facet Du, Hui-Peng
Ai, Yang
Jiang, Xiao-Hang
Zheng, Rui-Chen
Ling, Zhen-Hua
contents This paper targets a new scenario that integrates speech separation with speech compression, aiming to disentangle multiple speakers while producing discrete representations for efficient transmission or storage, with applications in online meetings and dialogue archiving. To address this scenario, we propose CodeSep, a codec-driven model that jointly performs speech separation and low-bitrate compression. CodeSep comprises a residual vector quantizer (RVQ)-based plain neural speech codec, a base-token disentanglement (BTD) module, and parallel auxiliary-token serial prediction (ATSP) modules. The BTD module disentangles mixed-speech mel-spectrograms into base tokens for each speaker, which are then refined by ATSP modules to serially predict auxiliary tokens, and finally, all tokens are decoded to reconstruct separated waveforms through the codec decoder. During training, the codec's RVQ provides supervision with permutation-invariant and teacher-forcing-based cross-entropy losses. As only base tokens are transmitted or stored, CodeSep achieves low-bitrate compression. Experimental results show that CodeSep attains satisfactory separation performance at only 1 kbps compared with baseline methods.
format Preprint
id arxiv_https___arxiv_org_abs_2601_12757
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle CodeSep: Low-Bitrate Codec-Driven Speech Separation with Base-Token Disentanglement and Auxiliary-Token Serial Prediction
Du, Hui-Peng
Ai, Yang
Jiang, Xiao-Hang
Zheng, Rui-Chen
Ling, Zhen-Hua
Audio and Speech Processing
This paper targets a new scenario that integrates speech separation with speech compression, aiming to disentangle multiple speakers while producing discrete representations for efficient transmission or storage, with applications in online meetings and dialogue archiving. To address this scenario, we propose CodeSep, a codec-driven model that jointly performs speech separation and low-bitrate compression. CodeSep comprises a residual vector quantizer (RVQ)-based plain neural speech codec, a base-token disentanglement (BTD) module, and parallel auxiliary-token serial prediction (ATSP) modules. The BTD module disentangles mixed-speech mel-spectrograms into base tokens for each speaker, which are then refined by ATSP modules to serially predict auxiliary tokens, and finally, all tokens are decoded to reconstruct separated waveforms through the codec decoder. During training, the codec's RVQ provides supervision with permutation-invariant and teacher-forcing-based cross-entropy losses. As only base tokens are transmitted or stored, CodeSep achieves low-bitrate compression. Experimental results show that CodeSep attains satisfactory separation performance at only 1 kbps compared with baseline methods.
title CodeSep: Low-Bitrate Codec-Driven Speech Separation with Base-Token Disentanglement and Auxiliary-Token Serial Prediction
topic Audio and Speech Processing
url https://arxiv.org/abs/2601.12757