Indirect scaling methods applied to the identification and quantification of auditory attributes

Auditory attributes, like for example loudness, pitch, sharpness, or tonal prominence, reflect how human listeners perceive their acoustical environment. The identification and of relevant auditory attributes and their quantification are therefore of major concern for different applications of sound quality research: be it in product-sound design or in the assessment of sound reproduction systems. In this Ph.D. thesis, so-called indirect scaling methods were developed and analyzed for their usefulness in sound quality evaluation. Such methods are characterized by requiring only simple qualitative judgments from listeners (as opposed to complex, often numerical judgments in direct scaling procedures), and by providing built-in tests of the validity of the underlying theoretical construct (as opposed to assuming its validity). In indirect scaling procedures, the numerical scale or the representation of the attributes are derived from modeling the listeners' judgments. The applicability of the developed methods was investigated in a series of experiments which aimed at identifying and quantifying auditory attributes of home-audio reproduction formats (mono, stereo, and multichannel formats). The results suggest that these methods have the potential to reveal aspects of auditory perception which cannot easily be obtained using existing direct scaling procedures.