An objective estimate of the perceived quality of reproduced sound in normal and impaired hearing

A new method for the objective estimation of the quality of reproduced sound for both normal-hearing and hearing-impaired listeners is presented. It is based on three parts: 1) Subjective sound quality ratings, 2) An auditory model, coupled to 3) An artificial neural network. The paper presents sound quality predictions on two perceptual scales; Clearness and Sharpness, and compares these to actual subjective ratings. These two scales were shown to be the most relevant for assessment of sound quality, and they were interpreted the same way by both normal-hearing and hearing-impaired listeners. The scales were found not to be absolute, thus the objective method cannot predict the absolute sound quality, but it can be used to rank the sound quality. Using test data from the present subjective rating experiment, the prediction error was found to be only slightly larger than the random variance in the subjective ratings.