Updated: February 24, 2006

Investigators:

Corine Bickley, Ph.D.
Anne Arrazcaeta

Purpose of this study:

To create a synthetic speech version of the City University of New York Nonsense Syllable Test; to compare performance on the synthetic test material with performance on the standard recording; and to compare the pattern of errors on the synthetic speech version to that obtained with the standard recording.

Project Description:

In our project, we will create a set of nonsense syllable materials that sounds as intelligible and natural as recorded speech and can be used for testing speech recognition performance. Previous work by the investigator seems to support the potential of such an approach (Bickley and Kidd, 1995). This type of approach could also be useful in assessment of performance with cochlear implants. As Franck, Xu, and Pfingst (2002) report, “It is crucial to understand the source of speech perception variability so that controllable variables can be adjusted to improve performance with the device.” The purpose of the proposed research is to determine whether synthesized materials, such as nonsense syllables, will sound natural and will provide an accurate measurement of phoneme recognition performance.

Project Progress:

To date six subjects have been tested. One listener has normal hearing, four have moderate, severe, or profound losses, and one has a mild low-frequency hearing loss. Stimuli (CV and VC syllables) were synthesized with HLsyn and presented in quiet and in noise in an audiological testing booth via supra-aural headphones. Results for voiceless stops and fricatives with the vowel “a” have been collected.

Presentation:

Arrazcaeta, A. & Bickley, C. Synthesized Nonsense Syllables for Hearing Aid Research using HLsyn 148th. Meeting of the Acoustical Society of America, San Diego, CA. Fall, 2004

For more information, contact: Corine Bickley, Ph.D.