Pitch perception

For many decades, pitch has been underrated; firstly, regarding its importance in auditory communication; and secondly, regarding the difficulties of understanding its perception.

Where auditory communication is concerned, it was of course throughout acknowledged that in music pitch is the most important "carrier of information". However, little attention was paid to the possibility that also in "ordinary life", i.e., in speech communication, and in auditory analysis and recognition of the huge variety of sounds impinging on our ears in daily life, pitch could be a key element. While it is apparent that in speech communication pitch is important for telling a woman's voice from a man's, it was little acknowledged that pitch might be involved far beyond that aspect. Although it has been known since Centuries (at least) that the ear segregates any complex sound into simple-tone pitches (spectral pitches), this phenomenon was more or less regarded as a kind of curiosity which was of interest only to theorists of music (such as, e.g., Tartini, and Helmholtz). By contrast, one can show that pitch is a most important carrier of information in a general, comprehensive sense. In its two essential varieties, i.e., spectral pitch, and virtual pitch, it plays the same role in auditory acquisition of information as is more or less obvious for primary contour and virtual contour in vision [75], [84], [87], [88], [93], [94], [96], [104], p. 21-30; cf. Hartmann (1988a, 1996a).

What regards understanding the perception of pitch, its difficulties and complexities have long been obscured by its seeming simplicity. The simple-minded version of pitch understanding is this:

Pitch is an auditory attribute of tones, and any tone is physically characterized by a frequency (i.e., oscillations per second). As everyone knows, pitch is immediately dependent on frequency, such that pitch ascends and descends monotonically with frequency. To build a model of pitch perception one merely has to design a frequency meter for tones.

However, careful analysis of the problem reveals that little is correct of this view.

Firstly, it is disregarded by this view that even single, isolated tones ordinarly have multiple pitches (except sine tones, see topic definition of pitch). So, the presupposition that throughout there is a simple one-to-one relationship between pitch and frequency, is wrong.

Secondly, it is disregarded by the simple-minded view that pitch is an auditory attribute not only of single, isolated tones, but also of tones that are accompanied by additional sound, and also of multiple simultaneous tones (sounds). In real life it is rare that one is exposed to single tones that are totally isolated from additional sound. Ordinarily, additional sound of any kind may be superimposed, and frequently multiple tones are sounding simultaneously. As a consequence, a model of pitch perception must a priori include a principle for the segregation of tones from noise, and for the segregation of multiple tones.

Thirdly, it is not true that any tone can be physically characterized by just one frequency, i.e., a well defined oscillation frequency. While for steady tones of many music instruments this is correct to a reasonable extent, it is not true for tones emanating from freely oscillating bodys such as, e.g., the strings of the piano, guitar, etc., and bells (see topics octave stretch, strike note of bells).

These notions render the simple-minded concept of pitch perception obsolete. A much more sophisticated approach to the perception of pitch must be pursued. Such an approach, i.e., the so-called virtual-pitch theory, is described in several of my publications, e.g. [17], [18], [22], [38], [55], [56], [104] Chap. 11. The benefit of such careful and comprehensive treatment of pitch perception is enormous. Certain fundamental questions are resolved, such as, e.g., the apparent conflict between the pitch of pure tones and that of complex tones, including the problem of the "missing fundamental" (see topics spectral pitch, virtual pitch, diplacusis binauralis). A considerable number of basic auditory phenomena get readily explained, e.g., the affinity of tones, octave equivalence, octave stretch, pitch shifts, the root of musical chords, the "acoustic bass" of the pipe organ, the strike note of bells, and the stretch of the musical tone scale. Moreover, new insights into auditory phenomena such as absolute pitch, and the dominance of a certain spectral region are obtained, along with new insights into musical consonance, and into the nature of auditory acquisition of information in general [75], [87], [93], [96], [104].

Author: Ernst Terhardt terhardt@ei.tum.de - Feb. 17, 2000

main page