[Mark DeBellis[_2_]]

On Dec 24 2012, 12:16=A0pm, wrote:
On Monday, December 24, 2012 11:13:50 AM UTC-5, Audio_Empire wrote:
No they've been disqualified because NAB said that the only way to cond=
uct
such a test is to play but a few seconds of a selection before switchin=
g to
the other DUT where only a few seconds are played through it before swi=
tching
back to the first DUT again.

No, I have never said that. What I said was that if you wanted a DBT to b=
e as sensitive to differences as possible, that's the way you should do it.=
That's scientifically accepted.

Of course, you're perfectly free to conduct your DBTs any way you like. I=
f you use longer passages without immediate switching, and you get positive=
results with all appropriate controls in place (double-blind, level-matche=
d, statistical significance), then I and others will accept them as correct=
..

But if you do that, then I guarantee we'll be able to get the same positi=
ve results using a quick-switching methodology!

BTW, here's a good article on the fallacy of long-term listening tests:ht=
tp://www.nousaine.com/pdfs/Flying%20Blind.pdf

bob

Hi Bob,

We discussed some of these topics a few years ago and I learned a lot
from the discussion, thank you (and nice to see you're still here!).
But I *still* think that perceptual equivalence is not the same as
indistinguishability.

I would be interested in your thoughts about the following.

First, it seems to me that it's possible that there could be two
signals, say three minutes of music each, where I can't distinguish
one signal from the other when I compare them, switching back and
forth, but where, nonetheless, I get greater pleasure from listening
to the first one (in its entirety) than to the second. It might be
difficult to *compare* the pleasure of one to that of the other, but
nonetheless it seems possible to me that, in fact, I might derive
greater pleasure from one than from the other.

Second, consider the following hypothetical example. Two recorded
excerpts, A and B, are identical, except that A has some added
ultrasonic component that, over short spans of time, causes a
temporary reduction in loudness sensitivity. Plausibly, the way A
sounds to the listener will not be the same as the way B sounds,
because the end of A will not have the same perceived loudness that
the end of B will have. However, it's not going to be easy to test
for this simply by comparing the two excerpts. If the listener
switches back and forth, the excerpts won't sound different, because
any reduction in sensitivity will affect the two equally. And if the
listener hears one excerpt in its entirety and then the other, he/she
has the problem of comparing stimuli that are distant in time, which
requires memory, which is not necessarily reliable.

The second example, which I am granting is hypothetical, would be one
where we would want to say (I think) that the way A sounds is not the
same as the way B sounds, though they might well be indistinguishable
in the relevant testing situations.

What do you think? Thanks.

Mark