Second Biennial Hearing Aid Research and Development Conference

September 22-24, 1997
National Institutes of Health
Bethesda, Maryland

Simulation of Reduced Dynamic Range with Recruitment of Loudness

The goal was to simulate reduced dynamic range, with recruitment of loudness, and to measure its effects on phoneme recognition in normally hearing subjects. Speech was filtered into three bands. Within each band, the rms. amplitude envelope was extracted from the speech waveform. A threshold was set, x dB below the maximum, and the range from x to maximum was expanded to cover the range 0 to maximum. This ìamplified center-clippedî envelope was then reapplied to the original waveform. The threshold, x, was varied from 3 dB to 30 dB. Phoneme recognition in consonant-vowel-consonant words was measured in 8 normally hearing subjects as a function of the threshold, x. The slope of the function relating performance to the value of x was similar to that obtained when performance is measured as a function of signal to noise ratio. This finding implies that the amplitude expansion (simulating recruitment of loudness) had no deleterious effect on phoneme recognition beyond that of the loss of low-intensity cues in the acoustic signal. In a second experiment, high-frequency emphasis was introduced before processing. This had the effect of increasing the slope of the performance versus threshold function and, consequently, of improving performance for subjects listening with very low dynamic range. An unexpected finding was differences in individual performance - not in the slope of the performance versus threshold function, but in its starting point. This finding indicates that even persons with normal hearing have different abilities to recognize phonemes on the basis of limited acoustic cues. Simulations of this type have potential value in the preclinical evaluation of novel processing strategies for hearing aids.

[Research supported by NIDCD]

Return to Conference Abstracts List

Top