Abstract:

In this work, techniques are developed and studied for the extraction of single-source acoustic signals out of multi-source cacaphony. Such extracted signals can be used in a variety of applications including: automatic speech recognition, teleconferencing, and robot auditory systems. Typical approaches fall into two categories: auditory modeling and array signal processing. The approach taken here is to combine these complementary techniques into an integrated one: auditory scene analysis-constrained array processing. Theoretically, this integrated approach should provide a performance gain since the information used by array processing, i.e., direction of propagation through a wave field, is independent of the signal structure information used by auditory scene analysis. One difficulty encountered by auditory scene analysis that can be overcome by array processing is the sequential grouping of spectrally dissimilar phonemes in a speech signal, such as a fricative followed by a vowel, or nasal.