Discrimination / Identification

[KEVIN AUSTIN <KAUSTIN@xxxxxxxxxxxxxxxxx>]

Having no scientific training, I was caught by Pierre Divenyi's comment
that ...
[snip]

> I would venture
>pointing out that (as they have shown) identification and discrimination
>can be regarded as esentially identical processes.


I have (intuitively) come to the place of finding the following, not
clearly segregated, sequence in learning (training), particularly in
ear-training
    acknowledgement (of source)
    discrimination
    ... [ reproduction ]
    identification
    ... [ notation ]

If I understand correctly, this would mean (something like, eg the
identification of intervals, or pianos), that there are / would
be a set of parameters that the listener has established to discriminate
that two (or more) stimuli are different.

There are likely sub-steps to this process in the identification of
intervals, starting with: are these two notes (piano sounds) the same.
This is made more complex with intervals if the two sounds come from two
different instruments (eg flute and horn, sampled Yamaha 9' and a
different Yamaha 9').

Identification would seem to require some sort of mental template to have
been established, that can reliably be called upon to compare the stimulus
to. (eg with intervals, what is the sound of a major 3rd ascending).

If this ability involves the same process as identification, then it
would seem likely that many (most?) people would be able to move from
discrimination to identification through training.

It has been my experience (with liberal arts music students), that there
are some people who never acquire the skill of interval (or chord!)
identifiction in a reliable, repeatable, long-term manner. Frequently this
is attributed to poor memory/recall.

A completely unscientific survey of mine (accumulated experience) informs
me that frequently, context plays as very large part in this process of
identification. Intervals (for most people that I have taught) appear not
to be two notes sounding together, but a collection of data related to
(western european) concepts of tonality, ie a set of relationships
between objects, rather than the objects themselves.

This may relate to the 'piano discrimination' issue, since few people
have listened intently to single notes on a piano, and likely even fewer
have more than 20% of this experience from in front of an acoustic
instrument.

If the subjects have listened to any pop music, over a variety of
loudspeakers (eg car, home, headphones, studio, kitchen), there may not
be a sense of the 'identity' of a piano. And this varies with how a piano
was recorded, and the processing that the sound underwent.

Acoustic pianos also have resonances which are always present (artifacts
of identity), notably that the dampers do not entirely mute all vibration
from the lower strings, and that the upper strings are undamped.

It could be that discrimination will work on the basis of the artifacts,
but the recognition/identification works on other information (but, as
proposed, through the same process).

Best

Kevin
kaustin@vax2.concordia.ca

Email to AUDITORY should now be sent to AUDITORY@lists.mcgill.ca
LISTSERV commands should be sent to listserv@lists.mcgill.ca
Information is available on the WEB at http://www.mcgill.ca/cc/listserv