Effects of multi-channel compression time constants on subjectively perceived sound quality and speech intelligibility.

OBJECTIVE: The purpose of this study was to determine the influence of the compression time constants in a multi-channel compression hearing aid on both subjectively assessed speech intelligibility and sound quality in realistic binaural acoustical situations for normal-hearing and hearing-impaired listeners.
DESIGN: A nonlinear hearing aid with 15 independent compression channels of approximated critical bandwidth was simulated on a personal computer. Various everyday life situations containing different sounds such as speech and speech in noise were recorded binaurally through original hearing aid microphones placed in BTE hearing aid cases. Two experiments were run with normal hearing and hearing-impaired subjects. For each subject, hearing thresholds were established using in situ audiometry. The static I/O-curve parameters in all channels of the hearing aid were then adjusted so that normal speech received an insertion gain corresponding to the NAL-R formula (Byrne & Dillon, 1986). The compression ratio was kept constant at 2.1:1. In the first experiment with six normal-hearing and six hearing-impaired subjects, the hearing aid was programmed to four different settings by changing only the compression time constants while all the parameters describing the static nonlinear Input/Output-curve were kept constant. The compression threshold was set to a very low value. In the second experiment with seven normal-hearing and eight hearing-impaired subjects, the hearing aid was programmed to four settings by changing the release time constants and the compression threshold while all other remaining parameters were kept constant. Using a complete A/B pair comparison procedure, subjects were presented binaurally with the amplified sounds and asked to subjectively assess their preference for each hearing aid setting with regards to speech intelligibility and sound quality. RESULTS AND
CONCLUSIONS: In Experiment 1, all subjects showed a significant preference for the longest release time (4 sec) over the two shorter release times (400 msec and 40 msec), pertaining to quality and intelligibility. In combination with the long release time, the attack time (1 msec or 100 msec) was of less importance. Larger inter-individual differences were observed among all subjects with regard to the perceived sound quality of musical and nonspeech signals. In Experiment 2 the hearing-impaired subjects showed a significant preference for the hearing aid setting with a long release time (4 sec) and a low compression threshold (20 dB SPL), both with regard to sound quality and speech intelligibility. With a short release time (40 msec), a lower compression threshold (20 dB SPL) was preferred over a higher threshold (50 dB SPL). Among normal-hearing subjects, the same settings were also preferred for speech intelligibility. However, no significant differences were observed in these subjects regarding the assessment of sound quality. Again, more individual variability was found during the assessment of music and nonspeech signals. In sum, the preference of a rather long release time raises the question of whether a compression system with a very short time constant is the optimal means to compensate for the sensory recruitment phenomenon.