Re: Segmentation (Al Bregman )

Subject: Re: Segmentation
From:    Al Bregman  <bregman(at)CCRMA.STANFORD.EDU>
Date:    Tue, 14 Sep 1993 16:25:09 -0700

From: Francesco Giomi <art(at)ifiidg.bitnet>
Subject:      Segmentation
X-To:         Auditory Mailing List <AUDITORY(at)VM1.MCGILL.CA>
To: Multiple recipients of list AUDITORY <AUDITORY(at)MCGILL1.BITNET>

I think that the problem is more similar to those founded
in electroacoustic music analysis than in tonal music.
The book of Lerdahl&Jackendoff concerns specifically with
segmentation of musical groups in tonal music and, however, could
be interesting to read for those problems.
But the problem of sound stream segmentation into objects is open
in electroacoustic music analysis. Perhaps you can find some ideas
about sonic objects and their perception in these two books:"
Chion, M. "Guide des object sonores" INA/Buchet/Chastel
Emmerson, S. (ed.)  "The language of electroacoustic music" (just
the article of Denis Smalley on spectromorphology).

Francesco Giomi
Musicological Dep. of CNUCE - Florence

DEAR FRANCESCO,
  THERE ARE PROBABLY PRIMITIVE "SEGMENTATION RULES" BASED ON THE PURE
ACOUSTICS OF THE SITUATION AND ALSO "PHRASE-GROUPING" RULES THAT MAY
BE MORE SPECTIFIC TO TONAL MUSIC.  AFTER ALL, NOTES ARE ALREADY
DISTINCT UNITS BECAUSE OF THEIR AUDIBLE ONSETS.  IT WOULD BE
IMPORTANT TO DISTINGUISH THESE, EVEN THOUGH THE LATTER MIGHT USE MANY
OF THE SAME FEATURES OF THE SOUND AS THE FORMER. I SUSPECT THAT THE
ONLY KIND OF RESEARCH CAPABLE OF DISTINGUISHING THEM WOULD BE STUDIES
OF HOW THE GROUPING TENDENCY CHANGES FOR CHILDREN AS THEY BECOME
FAMILIAR WITH MUSIC.  THIS COULD ONLY BE DONE WHERE THE TWO PRESUMED
PROCESSES MADE USE OF DIFFERENT FEATURES OF THE SIGNAL.

THANKS FOR YOUR REPLY.
 - AL

AL BREGMAN