Comments on Deutsch’s musical scale illusion

Abstract

Furthermore, the majority of the right-handed subjects reported that they heard Stream 1 in the right ear and Stream 2 in the left ear regardless of which channel was presented to which ear. I am presenting here an explanation of the tonal pattern organization aspects of Deutsch's illusion based on recently developed rules of auditory pattern perception and known principles of auditory localization.' This explanation is supported by the perceptual reports resulting from several manipulations of Deutsch's stimuli. I will not offer any explanation of the lateral asymmetries Deutsch reported. The perceptual rules on which this explanation is based are those of auditory stream segregation. When rapid tone sequences (about 1-20 tones/sec.) are presented in which successive tones vary widely along such dimensions as pitch (Bregman & Campbell, 1971), location of sound source (Judd, Note 1), loudness (van Noorden, 1975), and possibly timbre, attack, and rhythm (Judd, Note 2), listeners tend to perceive the tones as grouped along these dimensions, i.e., they may hear a high-pitched melody approximately simultaneously with a low-pitched melody, a melody on the left approximately simultaneouslywith a melody on the right, or a loud melody approximately simultaneously with a soft melody. This separation of a single sequence of notes into two or more perceived melodies or streams is called stream segregation (SS) (van Noorden's fission). The term is modified according to the dimension of variation, i.e., pitch stream segregation (PSS), spatial stream segregation (SSS), loudness stream segregation (LSS), etc. When the tone presentation rate is decreased and/or the separation along the dimension of variation is decreased, the sequence is heard as a unified whole, i.e., there is . pattern integration (van Noorden's temporal coherence). At intermediate speeds and separations, the stimulus can generally be perceived in either mode depending on the attentional bias of the listener and possibly other factors. This range of stimulus parameters is designated the ambiguous region. These relationships are summarized in Figure 1. (Note that the perceptual phenomena are actually more complicated for loudness; see van Noorden, 1975.) Deutsch's illusion involves a stimulus in which the information favorable to a perceptual organization based on pitch range is in conflict with information favorable to a perceptual organization based on the spatial separation of the sound sources. In other words, if the notes are heard as organized by pitch, a particular pair of melodies or streams should be