PHD WORK DONE

Objective
The main objective was to provide a hearing aid for hearing impaired. Another objective was to increase the spectral resolution in hearing aid which is simple and cost effective and provide binaural hearing aid with dichotic presentation of acoustic signal
Description
A hearing aid and method for compressing power spectrum and splitting an acoustic signal to make speech signal more audible for hearing impaired persons. An input acoustic signal is divided into a plurality of multiple spectral bands. A gain of each spectral band is derived from audiogram and adjusted. The plurality of modified spectral band is added as odd-index bands and even-index bands for binaural presentation of input acoustic signals. The power spectrum compression to frequency bands is performed during filtering. The present work is particularly applicable to hearing aid for hearing impaired with high frequency loss.
Conclusion
The processing scheme of modified wavelet packets for binaural dichotic presentation helped in improving the speech perception degraded due to increased masking. The processing scheme ps-mod and ps-Sym resulted in improvement in response time, recognition scores, and information transmission for consonantal features particularly for duration, frication and voicing features. Improvements are higher at higher levels of simulated loss. There was maximum improvement for duration feature. The processing scheme resulted in improvement in perception of place feature also. Improvement in perception of place feature can be attributed to reduction in the spectral distortions. Processing generally has no adverse effect on the reception of features such as manner, and nasality, and in most cases there is a moderate increase. All processing schemes were implemented as off-line processing and also real-time processing using DSP processors, the results for hearing impaired were nearly same for both processing.