Effects of Compression Release Time in Hearing Aid on Acoustic and Behavioural Measures of Speech

Abstract

The present study investigated the effects of varying the release time (40 ms, 640 ms & 1280 ms) on acoustic and behavioural measures of speech. The acoustic measures included a temporal measure or Envelope Difference Index (EDI) and a spectral measure or Log Likelihood Ratio (LLR). These were computed for VCV tokens of six classes of speech sounds. The behavioural measures were Speech Identification scores (SIS) and quality ratings for loudness, clarity, naturalness and overall impression. The effect of the release time was investigated at three input levels (30, 45 & 65 d BHL). The behavioural measures were performed on Group A participants (N=l0; with moderate to moderately severe HL) and Group B participants (N= 10; with moderately severe to severe HL). The acoustic measures were performed on five individuals (hearing loss above moderately severe) using the mean audiometric data of Group A and Group B. The effect of release time on EDI and LLR varied at each input level. At low input level of 30 dBHL. the effect of release time was not significant on all speech sounds. At 45 and 65 dBHL, shorter release time (40ms) showed significantly higher temporal and spectral distortions than longer release times (640 and 1280 ms) for majority of speech sounds. The amount of temporal and spectral distortions was found to be proportional to input level. SIS and quality ratings increased as the release time was made longer. Although not significant, Group B showed higher distortions and lesser SIS than Group A.