Sensory consonance

The term sensory consonance is meant to account for the auditory phenomenon that sounds of any kind in general differ with respect to how "pleasant", or "unannoying" they are to a listener [30], [31] [63]. Currently also the term sensory dissonance is often used, meaning the same phenomenon - just with the opposite sign. So, sensory consonance may also be defined by saying that it denotes the extent to which annoying factors are missing in a sound. Though the phenomenon and the concept of sensory consonance apply to any type of sound, it was originally designed as a component of musical consonance, i.e., with regard to the sounds of music.

While sensory consonance basically is dependent on three fundamental auditory attributes, namely, roughness, sharpness, and tonalness [37], [52] Aures (1985a, 1985b), Cardozo & van Lieshout (1981a), for the sounds of music and for their musical consonance, roughness is most important. This is consistent with the concept designed by Helmholtz (1863a), who termed the same phenomenon consonance (Konsonanz), and pointed out that it is essentially governed by auditory roughness.

With regard to musical consonance, sensory consonance was particularly considered for musical chords and dyads. For dyads, the occurrence of beats (and such, auditory fluctuation sensation, and roughness) intimately depends on the frequency interval, and - for harmonic complex tones - on the frequency ratio of the two tones forming the dyad. In a number of experiments by several authors, listeners were asked to estimate the degree of consonance of dyads composed of either sine tones or harmonic complex tones (Guthrie & Morrill 1928a, Plomp & Levelt 1965a, [34]).

According to the results of these experiments, both for dyads of sine tones and of harmonic complex tones, estimated consonance typically drops to a pronounced minimum when the interval is increased from unison to about 1.5 semitones. When the interval is enlarged beyond that, estimated consonance recovers continuously with interval width. For dyads of sine tones, the curve depicting sensory consonance as a function of interval width is smooth, i.e., it does not show any maxima at the special intervals that in music theory are regarded as "consonances", i.e., the third, fourth, fifth, and sixth intervals. For dyads of harmonic complex tones such intermediate maxima were observed (Plomp & Levelt 1965a); however, they are so little pronounced that, e.g., the fifth is estimated less consonant than the major seventh [104] p. 402.

As these results were shown to be not appreciably different when the experiments were carried out with musically trained listeners as opposed to musically naive listeners, the first important conclusion is that even when asked to estimate "consonance", listeners do not estimate consonance in the sense of music theory, but do estimate sensory consonance, where sensory consonance is something different from affinity of tones.The second conclusion then is, that tone affinity must be a distinctly more subtle phenomenon than sensory consonance, because, obviously, in the experimental competition between tone affinity and sensory consonance, the latter dominates.

The third conclusion is that in the above type of experiment sensory consonance is essentially governed by auditory roughness (or, in more general terms, any auditory attribute pertinent to beats). We have experimentally verified, that, indeed, there is an inverse relationship between estimated consonance and estimated roughness of dyads of sine tones [34], [104] p. 403.

As a more general conclusion one may say that the harmonic value and harmonic function of a musical dyad or chord is largely decoupled from occurrence of beats, i.e., roughness. While a certain amount of roughness indeed may go along with any sound of music, that roughness is of secondary significance where the musical function of the sound is concerned. This can be readily observed in real music. For instance, a dyad in the fifth interval produces a sound with a pronounced roughness when played on the piano in the bass region (e.g., C2-G2); it becomes "pure" (i.e., "sensorily consonant") only in a higher register. When instead of the fifth, e.g., a minor fifth is played in the bass region (C2-Gb2), its roughness is about the same as that of the fifth. So there is indeed no correlation between roughness and "theoretical" consonance. Composers obviously are little concerned about the roughness that occurs from chords in the bass region. Of course, also in the bass region chords are designed to serve a functional purpose, and roughness is either just tolerated or even intended as an expressive effect.

Vice versa, in experimental tests on the "consonance" of chords and dyads of tones with inharmonic partials it turned out that, again, it was the occurrence of beats and roughness that governed the results (Pierce 1966a, Slaymaker 1970a, Mathews & Pierce 1980a, Geary 1980a). That is, maximal estimated consonance was obtained for frequency intervals that by conventional music theory are regarded as disharmonic. These intervals were estimated as relatively most consonant because for them roughness was minimal.

Summarizing the role of sensory consonance in music, one can say that this phenomenon without a doubt is involved. However, the role which sensory consonance actually plays in music is far from determining the functional principles of harmony, on which tonal music is based. Sensory consonance, with its three major components roughness, sharpness, and tonalness, just accounts for the "sensory pleasantness" that is an aspect of any type of sound.

Author: Ernst Terhardt - Feb. 15, 2000

main page