Pitch salience in chords of harmonic complex tones

Abstract

Parncutt (1993) estimated the salience of the 12 chroma in chords of octave-complex tones (OCTs, Shepard tones). In each trial, listeners rated how well an OCT went with a preceding chord. The conventional root tended to go better than other chord tones. Non-chord pitches tended to go better if they corresponded to steps of associated scales; tones following the major triad went well if they belonged to the major scale (perfect fourth, major sixth, major seventh), following the minor if in (a) minor scale (major second, perfect fourth, minor sixth, minor seventh). The non-chord tones that went well tended to be missing fundamentals of incomplete harmonic series whose salience can be predicted by spectral or temporal models of pitch perception. Thus, it is unclear whether the data were determined by pitch perception, musical familiarity or both. We repeated the experiment using chords of harmonic complex tones - closer to real music. 20 Western musicians (students at the University of Graz) participated (the task was too difficult for non-musicians). 5 triads (major, minor, diminished, major third plus tritone, suspended fourth) were each presented in root position and two inversions. Each chord was followed by 12 probe tones. 180 trials were presented in a random order that differed for each listener. Chords were built from piano tones. Probe tones were OCTs and were slightly quieter than the chords. Each trial was randomly transposed. Listeners rated how well the tone went with the preceding chord on a 7-point scale. When averaged over inversions, results are similar to those of Parncutt (1993) for chords of OCTs. They are consistent with, but often do not confirm, the following assumptions: roots are more salient than other chord tones, non-chord pitches corresponding to missing fundamentals of incomplete harmonic series are more salient than other non-chord tones, and outer voices are more salient (as predicted by models of masking). The nature-nurture question remains unresolved. A possible interpretation is that pitches at missing fundamentals influenced the historical development of tonal-harmonic syntax, which in turn influenced preferences for specific chord progressions and, via frequent exposure, the perception of relationships between individual triads and preceding and following passages.