[Audio_Empire[_2_]]

In article ,
ScottW wrote:

On Wednesday, September 25, 2013 10:22:42 AM UTC-7, Scott wrote:
On Wednesday, September 25, 2013 9:50:39 AM UTC-7, Andrew Haley wrote:




With any test of audibility one has to implement controls for false
positives and false negatives. Bias controls against a false positive are
the easy part. Bias controls against false negatives can be a bit
trickier. Also one has to test the test for sensitivity. that is not so
easy either.

I never understood this point. If one doesn't think they hear a
difference...what is point of proving that they do?

For the individual, there would be no such point, but the purpose of a
bias controlled test is to obtain a statistically significant result as to
whether there actually exists audible differences or not. Humans are very
subjective creatures. We can talk ourselves (either consciously, or
subconsciously) into hearing - or not hearing - a myriad of things just
because we WANT to hear or not hear them. For instance, if a guy spends
$4K on a pair of interconnects to go between his preamp and his amp, believe
me, they ARE going to be the biggest improvement he ever experienced with
his system. Even if in a subsequent DBT he can't tell the difference between
his new $4K babies and a set of Radio-Shack $5 specials, he will swear that
the high-priced cables improved his system's sound. There are people
who post here regularly who have convinced themselves that everything
pretty much sounds the same - even when a DBT shows otherwise. But they
have painted themselves so deeply into that corner, that they REFUSE to hear
differences even when those differences are very easy for anyone to hear
because they are so gross (like with speakers).


I would suggest that a person who doesn't think they can hear a difference is
simply not a good subject for such a test.

Yes. That is true, but still, that's something that is hard to determine
beforehand.


Even if you don't believe that they can't hear a difference, what is the
point in trying to prove they do? All you could prove with great effort is
they're being deceptive which would only disqualify them from the
test....something you should have concluded when they said they can't hear a
difference.

Yes, but if I understand your question correctly, the way that properly
executed DBTs are designed (or at least the way I understand it), one
or two tin-eared listeners out of many and over many tries, won't, appreciably,
alter the results. Also, most people who take part in these tests probably
don't know beforehand whether or not that can hear a difference. Some, of
course, will go into such a test determined to NOT hear a difference even
when one exists. Statistically, their results won't make any difference to the
outcome either. The general wisdom with DBTs seems to be that one either
hears a difference between the devices under test or one doesn't. I'm
convinced that in most cases, a statistically positive result can probably
be very accurate, I'm less sure about a statistically negative result.

Also, consider the example of AEs claim of audible artifacts on all MP3s,
regardless of bit rate. I don't hear them on high quality VBR files. I
might with training be able to discern what he's talking about....but I'd
rather not.

I don't blame you there. I wish I didn't hear them, it would make things much
easier, but I do hear them. I have trained myself to be able to listen to Internet
radio as long as the data rate is 128 kbps or greater on SPEAKERS. I still can't
stand to listen on headphones. I convinced myself that Internet radio is
basically no more flawed than FM. It's just that the flaws are different. I.R. has
compression artifacts, and FM has multipath, dynamic range compression,
and hard limiting, plus, even full quieting on FM isn''t all that silent. I.R., OTOH,
is very quiet (at least) and the better feeds aren't compressed or hard limited.
That makes the digital compression artifacts more palatable.

ScottW