Online Spectrogram Inversion for Low-Latency Audio Source Separation

TL;DR

This paper introduces an online version of the MISI algorithm, called oMISI, which enables low-latency audio source separation with proven convergence and comparable performance to offline methods, suitable for real-time applications.

Contribution

The paper provides a rigorous optimization framework for MISI and develops oMISI, an online algorithm for real-time spectrogram inversion in audio source separation.

Findings

oMISI performs as well as offline MISI in speech separation tasks

The convergence of the proposed online algorithm is theoretically proven

oMISI enables low-latency, real-time audio source separation

Abstract

Audio source separation is usually achieved by estimating the short-time Fourier transform (STFT) magnitude of each source, and then applying a spectrogram inversion algorithm to retrieve time-domain signals. In particular, the multiple input spectrogram inversion (MISI) algorithm has been exploited successfully in several recent works. However, this algorithm suffers from two drawbacks, which we address in this paper. First, it has originally been introduced in a heuristic fashion: we propose here a rigorous optimization framework in which MISI is derived, thus proving the convergence of this algorithm. Besides, while MISI operates offline, we propose here an online version of MISI called oMISI, which is suitable for low-latency source separation, an important requirement for e.g., hearing aids applications. oMISI also allows one to use alternative phase initialization schemes exploiting the temporal structure of audio signals. Experiments conducted on a speech separation task show that oMISI performs as well as its offline counterpart, thus demonstrating its potential for real-time source separation.