Presented is a method for shaping the audio spectrum in order to improve speech intelligibility for the hearing impaired. Audiological data is used to design a finite impulse response (FIR) digital filter using the frequency sampling method. Cubic spline interpolation is used to generate any number of frequency samples from the limited number available from audiological data. The computer aided design procedure presented allows the filter designer to specify hearing threshold levels and threshold of discomfort levels at audiogram frequencies. The algorithm uses cubic spline interpolation to generate a smooth curve of gain versus frequency. Next, root-mean-square (RMS) error is calculated for a user specified range of FIR filter taps. This allows the designer to select a number of taps that will produce results in close agreement with the desired response. Finally, the frequency sampling method translates the data to a set of real impulse response coefficients for a digital filter. The frequency response of the FIR filter has linear phase with variable bandwidth. The set of real impulse response coefficients is stored in a file that is compatible with Interactive Laboratory System (ILS) software. ILS software is used to plot the frequency response of the filter and test the results by convolving the impulse response coefficients with digitized speech.
Master of Science (M.S.)
College of Engineering
Length of Campus-only Access
Masters Thesis (Open Access)
McLean, Brian D., "Spectral Shaping for the Hearing Impaired" (1987). Retrospective Theses and Dissertations. 5089.