"Steven Sullivan" wrote in message
...
William Sommerwerck wrote:
Last week at the Home Entertainment Show in New York Arny Krueger
participated in a panel discussion with John Atkinson, editor of
Stereophile
magazine. Arny is well known for his support for the scientific method
to
test what is audible and what is not.

The scientific method is not foolproof. Simply removing certain obvious
forms of
bias does not mean the test results are accurate or are correlated to
what we
"actually" hear when we sit down to listen.

That you think it *should* correlate to that, suggests you don't get why
blind tests are needed in the first place. What you 'actually' hear
when you sit down to listen is *NOT* a good reference point, when
differences
are 'actually' subtle or nonexistant.

This 'trust your ears' business that audiophiles tend to use as a mantra,
reflects a fundamental overestimation of how 'trustworthy' your ears
are, when they aren't allowed to be the *only* arbiters of what you are
hearing. What you 'actually' perceive when you sit down and listen in
casual
evulation, is an amalgam of truly audible plus other non-audible
'confounding'
factors. Science may not be foolproof, but the existnce of such factors
has been proved about as well as *anthing* has been. It's why scientific
investigations of all sorts routinely employs bias controls.
Cognitive/perceptual
confounding factors are *insidious* and *pervasive*.


Good science (as opposed to bad or pseudo-science) also pays excruciating
attention to the design and underlying premises/assumptions at work in the
test, to make sure that the scientist is measuring what he thinks he is
measuring. Arny and other DBT advocates have an almost-religious belief in
the efficacy of dbt's for any and everything audio..despite the huge
difference between measuring "sound" which is pretty much a physical
property, or "artifacts" which are discrete effects that one can train to
hear, and "music" which modern brain explorations have shown is hardwired in
some aspects into the brain and totally non-intuitive as to how things work.

Even the simple assumption that there are known thresholds that Arny and
Steven and others hold as "proof" that differences cannot exist if ABX
testing shows a null, now appears dubious as recent research suggest that
the brain "pre-conditions" the auditory nerves to focus on certain selective
affects depending on the context of what it is expecting and can exceed
previously thought thresholds in doing so (note that this is context
dependent and not likely to be operable in quick-switch "snippet" testing)..

Furthermore, open ended evaluation of equipment reproducing music doesn't
come with flags or signs saying "listen for this effect" or "catch how well
I handle this". The open-ended evaluative process requires the context of
the music itself and relaxed, unconscious exposure to allow the relevant
felicities or abrogation from what sounds "real" to emerge. Then also
factor in that psychophysiological research has show that the emotional
response triggered involuntarily by some aspects of music (and presumably
with music reproduction as well) do correlate with statistically significant
accuracy to higher "ratings" for the musical experience. And they take as
much as twenty seconds to build or disappear and only develop "in context".
Finally, factor in as well the recent finding that the ear nerves themselves
apparently have a "memory" for music apart from the remainder of the brain
such that they literally can "fill in the blanks" of music which is known,
even when the sound is physically cut off, and you can see how dubious a
simple dbt test becomes as a suitable test for open-ended evaluation of
equipment quality when reproducing music. Vastly different than listening
for known artifacts or broadband signal levels. A real scientist would be
asking more questions than ever today, and exploring the implications for
testing protocols, not promoting a "one-size-fits-all" solution and its
accompanying web site.