The role of temporal and spectral cues in the temporal integration of pitch and in pitch-based segregation of sound sources

Clark, Nicholas R (2011) The role of temporal and spectral cues in the temporal integration of pitch and in pitch-based segregation of sound sources. PhD thesis, University of Nottingham.

[thumbnail of CLARK_NR_PhD_Thesis.pdf]
Preview
PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (2MB) | Preview

Abstract

The auditory nerve conveys spectral information, reflecting the location of maximum vibration along the frequency-tuned basilar membrane, and also information reflecting the timing of peaks in the vibrations at each location. Debate continues as to whether pitch is extracted based on the available temporal or spectral representations of tonal stimuli, or both. The aim of the current work was to determine the roles of temporal and spectral harmonicity cues for pitch, under important conditions for understanding speech in multi-talker environments. Two such conditions are the temporal integration of pitch and pitch-based segregation of sound sources.

Pitch information in running speech changes over time. Therefore, pitch-extraction mechanisms must be able to follow these changes to enhance intelligibility, particularly when listening in modulated backgrounds such as competing speech. However, the temporal resolution of pitch has received little attention. In the first three chapters, the roles of temporal and spectral cues on the temporal resolution of pitch extraction were determined by measuring pitch-domain temporal modulation transfer functions and gap-detection thresholds. Temporal resolution was shown to be unaffected by the availability of spectral cues, and similarly unaffected by the overall pitch strength of the stimulus. However, the system was much more sluggish in response to changes in pitch information in stimuli presented in high-frequency regions compared to low-frequency regions. This processing strategy may reflect the progressive loss of accurate temporal information towards higher frequencies imposed by transduction processes in the auditory periphery.

To understand speech in noise, the ability of the auditory system to integrate pitch information over long periods is equally important as its ability to detect rapid changes in pitch. In Chapter 4, discrimination thresholds for pitch value and pitch strength were measured in the presence and absence of spectral cues as a function of stimulus duration. The assumption was that discrimination thresholds would reach asymptote at the stimulus duration corresponding to the length of the pitch integration window. However, the pitch-strength discrimination data revealed integration was only limited by the stimulus duration, suggesting that this task may reflect the rate of decrease in the variance of internal pitch-value and pitch-strength estimates with increasing stimulus duration, but not the total integration capacity of the system.

In multi-talker environments, listeners have to process multiple simultaneous tonal sound sources. The fifth study showed that temporal interactions between simultaneous tonal stimuli could aid detection in the absence of spectral cues. In contrast, harmonic resolvability is thought to be a prerequisite for pitch-based simultaneous grouping. However, data from a second experiment showed that listeners were able to perceptually segregate tonal sounds in the absence of spectral cues.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Krumbholz, K.
Crowe, J.A.
Subjects: Q Science > QP Physiology
Faculties/Schools: UK Campuses > Faculty of Engineering > Department of Electrical and Electronic Engineering
Item ID: 11959
Depositing User: EP, Services
Date Deposited: 25 Oct 2011 12:57
Last Modified: 16 Dec 2017 05:19
URI: https://eprints.nottingham.ac.uk/id/eprint/11959

Actions (Archive Staff Only)

Edit View Edit View