Music Source Separation with Band-Split RoPE Transformer ...

Music source separation (MSS) aims to separate a music recording into multiple musically distinct stems, such as vocals, bass, drums, and more. Recently, deep learning approaches such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs) have been used, but the improvement is...

Full description

Bibliographic Details
Main Authors: Lu, Wei-Tsung, Wang, Ju-Chiang, Kong, Qiuqiang, Hung, Yun-Ning
Format: Article in Journal/Newspaper
Language:unknown
Published: arXiv 2023
Subjects:
Online Access:https://dx.doi.org/10.48550/arxiv.2309.02612
https://arxiv.org/abs/2309.02612
Description
Summary:Music source separation (MSS) aims to separate a music recording into multiple musically distinct stems, such as vocals, bass, drums, and more. Recently, deep learning approaches such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs) have been used, but the improvement is still limited. In this paper, we propose a novel frequency-domain approach based on a Band-Split RoPE Transformer (called BS-RoFormer). BS-RoFormer relies on a band-split module to project the input complex spectrogram into subband-level representations, and then arranges a stack of hierarchical Transformers to model the inner-band as well as inter-band sequences for multi-band mask estimation. To facilitate training the model for MSS, we propose to use the Rotary Position Embedding (RoPE). The BS-RoFormer system trained on MUSDB18HQ and 500 extra songs ranked the first place in the MSS track of Sound Demixing Challenge (SDX23). Benchmarking a smaller version of BS-RoFormer on MUSDB18HQ, we achieve ... : This paper explains the SAMI-ByteDance MSS system submitted to Sound Demixing Challenge (SDX23) Music Separation Track. Version 2 of paper fixed some typos ...