Date of Award

Spring 1-1-2010

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Speech, Language & Hearing Sciences

First Advisor

Kathryn H. Arehart

Second Advisor

James M. Kates

Third Advisor

Gail Ramsberger

Abstract

Speech perception depends on access to spectral and temporal acoustic cues. Temporal cues include slowly-varying amplitude changes (temporal envelope) and higher-rate amplitude changes (temporal fine structure, TFS). This study sought to quantify the effects of alterations to temporal structure on the perception of speech quality by parametrically varying the amount of TFS available in specific frequency regions.

The three research aims were to: 1) establish the role of TFS in quality perception, 2) determine if the role of TFS in quality perception differs for listeners with normal hearing and listeners with hearing loss, and 3) quantify the relationship between intelligibility scores and quality ratings. Quality ratings were obtained using an 11-point scale for three signal processing types (two types of vocoding noise and total band removal), and with different amounts of background noise (none, 18, and 12 dB signal-to-noise ratios (SNRs)) for a range of frequency regions.

TFS removal above 1500 Hz had a small, but measurable, effect on quality ratings for speech in quiet (i.e. a 2.2-point drop on an 11-point scale). For speech in noise, TFS removal had a smaller effect (at most a 1.2-point drop). TFS modifications also influenced the temporal envelope. Analyses using the Hearing Aid Speech Quality Index (HASQI) (Kates & Arehart, 2010) showed that temporal envelope modifications provide a partial, though incomplete, description of sound quality degradation. Thus, TFS is important to consider in models of quality perception of speech.

Intelligibility performance was correlated with quality ratings, with larger correlations evident for poorer intelligibility. However, a significant relationship between intelligibility and quality was documented even when intelligibility remained above 95%.

The results of this study have both scientific and clinical implications. The findings provide insight into the mechanisms that affect sound quality perception, including the role of TFS. Additionally, this knowledge may be applied to future signal processing design, helping to maximize both speech intelligibility and sound quality in new hearing aids.

Share

COinS