Sound characteristics of multi-tone sounds

Tonal sounds are included in many everyday sounds and they are known to influence the perceived loudness, the sound character and the evaluation of a sound. Especially for a high number of simultaneous tonal components the perceptual effects are astonishing and yet not fully understood. In order to reduce adverse effects of such sounds, it is necessary to identify the underlying perceptual aspects driving the unpleasantness and annoyance of multi-tone sounds. The overall aim of this PhD-thesis is the characterization and quantification of the sound character and the appraisal resulting from temporal and spectral characteristics of a particular type of multi-tone sounds. The sounds of interest consist of two complex tones and additional combination tones and they can be regarded as generic representatives of signals produced by two sound sources which are coupled by a non-linear interaction. After a first exploration of the perceptual space resulting from alterations of the temporal signature, the quantitative impact of import technically important signal parameters on the loudness judgments and the preference evaluations of the sounds were quantified with a special method. In comparison to one common reference sound, levels at the points of subjective equality (PSEs) for loudness and for preference were determined in separate experiments. The level difference between the preference and the loudness PSEs is used a quantitative measure for preference relevant aspects of the sound character which in return becomes tangible and quantifiable on a dB-scale. A prediction model for the PSE levels, based on only two identified descriptor variables, explains already 87 % of the variance. The fitted parameters of the model disclose the quantitative relationships between the preference equivalent dB(A)-level and the sound character of the multi-tone sounds, which is characterized by the sharpness (according to the DIN 45692 standard) and the repetition rate of the signals.