Abstract
Effects of frequency-compression in terms of the intelligibility in severe sensori-neural hearing loss were investigated by using a frequency-compression technique of speech signal with higher fidelity partial auto-correlation coefficient (PARCOR) synthesis method. The hearing level of cases were 70s dB in 1 ear, 80s dB in 6, 90s dB in 11, 100s dB in 8, and 110s dB in 1. The test sound, phonated by a female announcer, consisted of 50 two figure numbers. Intelligibility scores were obtained in every frequency-compressed rate which was 100% to 60% in 10% steps. Forty-one percent of the ears examined showed a 20% or better score at the proper rate of frequency-compressed sound than the score at non-compressed sound. Almost all cases showed a score between 10% to 60% at non-compressed sound, and the intelligibility was clearly impreved. The relation between the hearing level or speech articulation score and the improvement of the intelligibility by frequency-compressed sound was not clear.
