Study of Spectral and Temporal Effects in the perception of noise degraded speech

Cochlear Implants (CI) are surgically installed biomedical gadgets that give hearing to significantly hard of hearing individuals by direct electrical incitement of the auditory nerve. It has been recognized that auditory simulation models in cochlear implants improve the perception of speech significantly. For music and hearing perception in noisy environments, phase cue plays a major role but it is not being conveyed by the present day auditory models. This paper describes the results of a study on the spectro-temporal effects on the perception of English phonemes comprising vowels and consonants. An auditory simulation model is developed to extract and process the spectral and temporal features using Hilbert algorithm. Six English phonemes are presented to normal hearing listeners under various signal-to-noise ratios (SNR) and their speech intelligibility scores are computed. Significant improvement is achieved and hence this model should be helpful for enhancing the speech intelligibility of cochlear implant subjects.