Paired comparison judgments in low‐fidelity circuitry

Abstract

The effects of electroacoustic distortions on speech transmission through hearing aids has met with little agreement as to which characteristics are most detrimental to the understanding of speech. The purpose of this study was to examine the capacity of listeners to yield proximity (dissimilarity) matrices and preferences of hearing aids through the method of paired comparisons. Evaluative data consisted of pairwise judgements of 12 hearing aids. Fifty normal hearing adults representing five different acoustic competencies were used as listeners. These five competencies included (1) musicians; (2) audiologists; (3) engineers; (4) audio enthusiasts; and (5) noncategorical. Their task was to judge the dissimilarity between 78 pairs of hearing aid transduced spech and music on a nine‐point scale and to mark their hearing aid preference. Characteristics for each of the 12 hearing aids were determined for the following parameters: overall frequency response, fundamental frequency response, second, third and total harmonic distortion, phase distortion, transient distortion, intermodulation distortion, and internal noise. This perceptual data was analyzed using procedures of multidimensional scaling. INDSCAL was used to determine those electroacoustic characteristics germane to the decision process of judging quality in low fidelity circuitry. Correlation analysis determined the regression structure of the derived electroacoustic distortions on preference. Data will be discussed examining the individual and group differences including the precision and variance by which subjects judge low‐fidelity circuitry.