A multidimensional technique for sound quality assessment

One approach to improving sound quality is to create a preference map on th e basis of several acoustic parameters relevant to auditory perception. The map is derived from several stages of subjective testing, acoustic analysi s, and auditory modeling. The multidimensional scaling technique CLASCAL re veals common perceptual dimensions shared by sets of sounds samples, percep tual features specific to each sound, and the different subject classes amo ng listeners. The listeners are asked to judge the degree of dissimilarity of all pairs of sounds on a continuous scale. The analysis gives a perceptu al spatial representation of the sounds. From this analysis, acoustic and a uditory modelling analyses can be performed to determine the stimulus param eters that are strongly correlated with different perceptual dimensions and , where possible, with the specific features. The next stage in the analysi s involves determining the probability of one sound being preferred to anot her. An analysis of the data allows a projection of the structure of listen ers' preferences onto the physical parameter space underlying the previousl y determined multidimensional perceptual space. In many cases, it is found that the physical parameters having the most effect on the listeners' prefe rences are dependent on the set of stimuli being compared. Furthermore, whe n one stimulus parameter is kept constant across trials, this may alter the effects of other parameters on the listeners' preferences. Therefore conte xt effects must be taken into account in multidimensional sound quality ana lysis, particularly since the qualitative aspects of most sounds are clearl y multidimensional.