RE critical bandwith, Plomp has demonstrated (R. Plomp, "Aspects of
Tone Sensation", Academic Press, 1976) that the experimental
estimation of the critical bandwith can only be a rough
approximation, since the values found are dependent on the
experimental conditions. According to his research, the values may
differ by a factor of two.

The interesting aspect of the critical bandwith is its non-linear
behaviour along the register: from about 500 Hz upwards it's a
function of the frequency (roughly 1/5 frq); for frequencies below
500 Hz it begins to remain more or less constant. That's why some
musical intervals that are larger than the critical bandwith at
higher frequencies may fall within the critical bandwith in the lower
register.

See also R. Rasch & R. Plomp, "The Perception of Musical Tones",
1982, repr. in D. Deutsch, "The Psychology of Music". This paper and
the others in the book discuss many important aspects of the
psychoacoustics of pitch, timbre, timing, grouping mechanisms in
music, etc. including an article by Risset and Wessel: "Explorations
of Timbre by Analysis and Synthesis".

luis jure