Design of a lab model of a Digital Speech Processor for cochlear implant

A cochlear implant (CI) is an electronic device that restores partial hearing to a person who is profoundly deaf or severely hard of hearing. In the current work, the TMS320C6713 DSK is used to implement a lab model of the cochlear implant using the uniform band channel vocoder model. The model uses a real-time speech input and the outputs are used to excite point electrodes. As the implantation of electrodes in the inner ear is a surgical procedure, validation of the functioning of the CI model is done using acoustic simulation results. Acoustic CI simulations of cochlear implants are performed using a cepstral vocoder model with Mel-frequency cepstral coefficients. As the performance of each model depends on the perceptual quality of the synthesized speech, the acoustic CI models were used to conduct a listening test using normal hearing subjects for words from the LNT database. The average grade for the uniform band channel vocoder model is found to be 3.9 as against an average grade of 3.1 on a scale of 5 for the cepstral vocoder model.