A General Flexible Framework for the Handling of Prior Information in Audio Source Separation

Abstract

Abstract-Most of audio source separation methods are developed for a particular scenario characterized by the number of sources and channels and the characteristics of the sources and the mixing process. In this paper we introduce a general audio source separation framework based on a library of structured source models that enable the incorporation of prior knowledge about each source via user-specifiable constraints. While this framework generalizes several existing audio source separation methods, it also allows to imagine and implement new efficient methods that were not yet reported in the literature. We first introduce the framework by describing the model structure and constraints, explaining its generality, and summarizing its algorithmic implementation using a generalized expectation-maximization algorithm. Finally, we illustrate the above-mentioned capabilities of the framework by applying it in several new and existing configurations to different source separation problems. We have released a software tool named Flexible Audio Source Separation Toolbox (FASST) implementing a baseline version of the framework in Matlab.

Automatic music transcription: challenges and future directions

“…In the context of source separation, Ozerov et al [101] proposed a framework that enables the incorporation of prior knowledge about the number and types of sources, and the mixing model. The authors showed that by using prior information, a better separation could be achieved than with a completely blind system.…”

“…Another approach is to perform instrument separation and identification jointly, using for example signal model-based probabilistic inference in the score-informed case [67]. Furthermore, ideas and algorithms from the field of source separation can be utilised for AMT, especially regarding the exploitation of spatial information where available [41,101]. On the other hand, music transcription can be used to improve source separation.…”

A general variational Bayesian framework for robust feature extraction in multisource recordings

“…In this formulation, f is the frequency index, n the time frame index, i the channel index and j the source index. Note that this framework also works for diffuse or reverberated sources by modeling each source as a subspace spanned by several point sources [5]. We adopt a local Gaussian model [5] for the source coefficients.…”

“…Note that this framework also works for diffuse or reverberated sources by modeling each source as a subspace spanned by several point sources [5]. We adopt a local Gaussian model [5] for the source coefficients. We set a zero-mean complex Gaussian prior over the source coefficients s j,f n with variance v j,f n :…”

“…In this paper, we extend this approach to the general modeling framework for source separation recently introduced in [5]. This framework generalizes a wide class of existing source separation algorithms, including nonstationarity-based frequency-domain independent component analysis (FDICA) and single-or multi-channel nonnegative matrix factorization (NMF).…”