The auditory organization of speech and other sources in listeners and computational models

Abstract

Speech is typically perceived against a background of other sounds. Listeners are adept at extracting target sources from the acoustic mixture reaching the ears. The auditory scene analysis (ASA) account holds that this feat is the result of a two-stage process. In the ®rst-stage, sound is decomposed into collections of fragments in several dimensions. Subsequent processes of perceptual organization reassemble these fragments, based on cues indicating common source of origin which are interpreted in the light of prior experience. In this way, the decomposed auditory scene is processed to extract coherent evidence for one or more sources. Auditory scene analysis in listeners has been studied for several decades and recent years have seen a steady accumulation of computational models of perceptual organization. The purpose of this review is to describe the evidence for the nature of auditory organization in listeners and to explore the computational models which have been motivated by such evidence. The primary focus is on speech rather than on sources such as polyphonic music or non-speech ambient backgrounds, although all these domains are equally amenable to auditory organization. The review includes a discussion of the relationship between auditory scene analysis and alternative approaches to sound source segregation. Ó

Spectral Clustering for Speech Separation

“…We base our design on classical cues suggested from studies of perceptual grouping (Cooke and Ellis, 2001). Our basic representation is a "feature map," a two-dimensional representation that has the same layout as the spectrogram.…”

Audio video based fast fixed-point independent vector analysis for multisource separation in a room environment

“…Solving the machine cocktail party problem requires the design of a method to focus on the desired speech signal while suppressing or ignoring all the other competing speech sounds [3]. Attempts to solve the machine cocktail party problem have come from the signal processing community in the form of blind source separation (BSS) [4] and generally from the computer science *Correspondence: Y.Liang2@lboro.ac.uk School of Electronic, Electrical and System Engineering, Loughborough University, Leicestershire, LE11 3TU, UK community in the form of computational auditory scene analysis (CASA) [5]. CASA is motivated by understanding the human auditory scene analysis.…”

Exploiting correlogram structure for robust speech recognition with multiple speech sources

“…blind source separation by independent component analysis (Parra and Spence, 2000), but separating and recognising speech in single-channel signals, the problem considered in this article, still remains a challenging problem. Human listeners, however, are adept at recognising target speech in such noisy conditions, making use of cues such as pitch continuity, spacial location, and speaking rate (Cooke and Ellis, 2001). They are able to effectively extract target audio streams from monaural acoustic mixtures with little effort, e.g.…”

“…listening to speech/music mixtures on a mono radio program. It is believed that there are processes in the auditory system that segregate the acoustic evidence into perceptual streams based on their characteristics, allowing listeners to selectively attend to whatever stream is of interest at the time (Bregman, 1990;Cooke and Ellis, 2001). This offers an alternative to techniques which require the noise to be effectively removed from the speech, e.g.…”