Comparison of hearing loss compensation algorithms using speech intelligibility measures

Sensorineural hearing loss includes loss of high‐frequency sensitivity which results in decreased speech intelligibility. The loss cannot be compensated by inverting the audiogram because of the non‐linear effects of sensorineural hearing loss (frequency smearing, decreased dynamic range, decreased time‐frequency resolution). Several non‐linear compensation schemes exist (Half‐gain, POGO, NAL‐R, Fig. 6, DSL and LGOB) and this paper provides a comparison of those using the objective Perceptual Evaluation of Subjective Quality (PESQ) score and the subjective Hearing In Noise Test (HINT). The listening tests were run on 15 unaided hearing impaired listeners as well as 15 normal hearing listeners using a simulated hearing loss algorithm. These results show marked improvement in intelligibility for the compensated speech over the normal speech for both normal and hearing impaired adults.