Auditory Grouping and Attention to Speech

The human speech recognition system is superior to machine recognition in many ways, but one of the most dramatic is in its resistance to additive "noise" [1]. The human listener is able to use a variety of types of information to help segregate the sounds from different sound sources. The pioneering work on the basic phenomena of human sound segregation was carried out by Al Bregman, and is reported in his book "Auditory Scene Analysis" [2] together with a theoretical analysis of the problem and types of solution. Bregman's own work has been mainly concerned with simple musical sounds; our own work [3, 4] has concentrated on exploring the relevance of auditory scene analysis to speech perception. This paper will briefly review some of the results from this work that may have implications for the machine recognition of speech.