Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site ur-tut.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!rochester!ur-tut!tuba
From: tuba@ur-tut.UUCP (Jon Krueger)
Newsgroups: net.audio
Subject: Re: Re: nut.audio:  The "ear" vs. the "instrument"
Message-ID: <155@ur-tut.UUCP>
Date: Wed, 9-Oct-85 16:33:40 EDT
Article-I.D.: ur-tut.155
Posted: Wed Oct  9 16:33:40 1985
Date-Received: Sat, 12-Oct-85 05:29:30 EDT
References: <787@nmtvax.UUCP> <569@edison.UUCP>
Reply-To: tuba@ur-tut.UUCP (Jon Krueger)
Organization: Univ. of Rochester Computing Center
Lines: 56


Yawn.  I can't believe no one has mentioned this yet: the ear is an
instrument.  Like any instrument, it has known limits, such as limited
accuracy, reliability, sensitivity, and so on.  So do instruments commonly
found in the lab.  Both are instruments for detecting things.

Whether my ears or your ears or golden ears can beat a given lab instrument
in sensitivity is a testable question.  Are there ears somewhere that beat
some piece of lab equipment in detecting presence/absence of some aspect of
sound at some level under some conditions?  Of course.  Is there a lab
instrument that can detect something better than any ears?  Of course.  Do
any of these questions and their answers tell us when it's appropriate to
use ears versus lab instruments to detect things?  Of course not.
Sensitivity is desirable in instruments that detect things.  But sensitivity
is useless without reliability.  It's common to find, in both ears and lab
instruments, that you can stretch one at the expense of the other.
Double-blind testing and the like are methods for improving reliability of
that instrument, the ear.

For more information on this view of the world, I recommend Green and Swets
"Signal Detection Theory and Psychophics".  But I don't think you have to
read the book to put this "ear vs. instrument" topic to bed.  My point is:
claiming that the ear is more sensitive than the lab instrument doesn't help
anyone.  It's got to be reliable too.  If you can turn an ear into a
reliable instrument by putting it together with a bunch of other ears
attached to a bunch of people who are participating in a double-blind
testing procedure, I'm all for it.  If you find ANY ears more sensitive to
something than the current crop of lab instruments, and you have ANY way of
making the ears reliable detectors, I'm all for it.  If you can't show
reliability, please don't flame about sensitivity.

I'm equally unimpressed by a lab instrument that someone claims can detect
things that nothing else can, but I can't rely on it to give the same report
to the same input every time I present the input, or I can't rely on it
to agree with another instrument like it every time we present them
both with the same input.  Fair enough?

As a final note, both ears and lab instruments are capable of generating
data on nominal or ordinal scales.  I know of no way of calibrating an
ear so as to get data out of it on an interval scale.  So if you want
to make statements like "CD Brand X is better than CD Brand Y" either
ears or lab instruments are appropriate detectors.  But if you want to
make statements like "CD Brand X is 2.5 times better than CD Brand Y"
(which might be compared to their relative prices and plugged into
a buying decision) you want to use lab instruments.

-- 

-- Jon Krueger
UUCP:	...seismo!rochester!ur-tut!tuba
BITNET:	TUBA@UORDBV
USMAIL:	University of Rochester
	Taylor Hall
	Rocheseter, NY  14627
	(716) 275-2811
"A Vote for Barry is a Vote for Fun"