Effective FPGA Implementation of Comb Filters to Improve Perception of Sensorineural Hearing Impaired

Abstract

Abstract

In this paper, effective method for implementation of comb filter on FPGA platform has been proposed. This work is intended for hearing aid to be used for people suffering from sensorineural hearing loss. The algorithmic implementation on FPGA is intricate during the design process. Here we propose effective and simplified approach for implementation of comb filter with 512 coefficients using Spartan-6 FPGA for dichotic presentation. The design of comb filter uses spectral splitting based on auditory critical bandwidths. Magnitude response of these filters are optimized for low perceived spectral distortion having high stop band attenuation and low pass band ripples. This method was evaluated by performing listening tests experimentation on hearing impaired people having different degree of sensorineural hearing loss. The speech material used consists of fifteen vowel consonants vowel syllables. The performances of the listening test have been analyzed with qualitative assessment, response time, recognition score and information transmission analysis. Results indicate significant reduction of response time, improvement in recognition score and transmission of all consonantal features. The latency was analyzed and found to be 8x10^-3 seconds (average), noticeable to subjects.

Keywords: Sensorineural Hearing Loss, Comb filter, FPGA, Dichotic Presentation, Latency

Cite this Article

Chopade JJ, Futane NP, Kharate GK. Effective FPGA Implementation of Comb Filters to Improve Perception of Sensorineural Hearing Impaired. Current Trends in Signal Processing. 2017; 7(3): 15–29p.