Interaction between Pitch and Timbre Cues on Auditory Stream Segregation and Grouping
Caitlin Salvagno
Yonghee Oh, PhD
Abstract
Problem Statement: In multi-talker listening environments, the culmination of different voice streams may lead to the distortion of individual messages of each source leading to deficits in comprehension. Voice characteristics such as pitch and timbre are major dimensions of auditory perception and play a vital role in grouping and segregating incoming sounds based on their acoustic properties. The current study investigated how pitch and timbre cues (determined by fundamental frequency and spectral tilt, respectively) can affect perceptual integration and segregation of complex-tone sequences within an auditory streaming paradigm.
Methods: Twenty normal-hearing listeners participated in a traditional auditory streaming experiment using two alternating sequences of harmonic complex tones A and B with manipulating fundamental frequencies and spectral tilts. Grouping ranges, the fundamental frequency ranges over which auditory grouping occurs, were measured with various spectral tilt differences between tones A and B (0, -0.5, and -1 dB/octave).
Results: Average results show that the grouping ranges decreased by 1.5 to 2 times as spectral tilt differences increased from 0 to -1 dB/octave.
Conclusions: Changing spectral tilts (timbre cue) between tones A and B leads to smaller grouping ranges. This concept is representative of culminating voices observed in the “cocktail party problem” and suggests timbre cues play a role in how we perceive the combination of simultaneous human voice pitches. Further investigation is needed with larger differences in spectral tilt as well as for hearing impaired listeners and cochlear implant users.