Re: maximum tatum (one tatum, two tata) (Pierre Divenyi )

Subject: Re: maximum tatum (one tatum, two tata)
From:    Pierre Divenyi  <pdivenyi(at)EBIRE.ORG>
Date:    Tue, 9 Apr 2002 17:31:40 -0700

I think that there should be an agreement between the tatumites as to what we are talking about. I presume the point of interest is perception, rather than production. Sticking to piano, one further note (no pun intended) is the enormous difficulty of achieving ensemble playing for two pianists, whether on one piano or two, mainly because the slightest inattention will result in "broken chords", i.e., in places when the two "were not together". The glissando example was an easy metric for demonstrating just how little successiveness between two temporally adjacent piano notes sounds like two piani. Psychoacoustically speaking, the spectral commonalities between successive notes, due to onset spectral splatter, can be regarded as an insurance that, when judging simultaneity or succession, the listener can use envelope analysis within similar spectral bands and thereby gain maximum precision (see Divenyi and Danner, P&P 1977, Divenyi and Sachs, P&P 1978). That is, the same game works even for random-frequency notes, because there is substantial energy in a given spectral band (there are several in piano notes) for the listener to perform effective envelope judgments. If one disregards the spectral splatter (say, by using sinusoidal tones with gradual onsets and offsets), then the perceived time interval between adjacent tones will be affected by the musical interval between them (again, see the above references) and who knows at what time separations the listener will hear two tones as successive. Thank heaven, there is spectral splatter in the output of most musical instruments -- hence the need for an orchestra conductor who is paid (mostly overpaid) to ensure that entries are successive only when the score calls for such. Pierre At 08:01 PM 4/9/2002 -0400, mercuri(at)GRADIENT.CIS.UPENN.EDU wrote: >I beg to differ with Pierre but there should be a difference >between predictive sounds (such as a glissando, or even Morse code) >and totally random onsets. One of your contributors mentioned multiple >sounds, and there too, there is a matter of attention (cocktail party >effect) and illusion (alternating octaves) where there has been some >research, although not terribly much in the way of discrete event >counting. Diana Deutsch's work is of note. > >Rebecca Mercuri.

This message came from the mail archive
maintained by:
DAn Ellis <>
Electrical Engineering Dept., Columbia University