A Brief History of CD DBTs
On Friday, December 21, 2012 10:38:52 AM UTC-8, wrote:
On Friday, December 21, 2012 9:51:42 AM UTC-5, Audio_Empire wrote:
=20
On Thursday, December 20, 2012 6:46:30 PM UTC-8, wro=
te:=20
=20
On Thursday, December 20, 2012 6:37:24 PM UTC-5, Audio_Empire wrote:
=20
=20
=20
=20
=20
I still believe that the only way to do a DBT/ABX of something as=
=20
=20
subtly different as amplifiers should be done with each amp being=
=20
=20
auditioned for as much as a half hour before switching to the other=
=20
=20
amp. Use the same cuts from test CDs for each session, played in=20
=20
the same order. Of course careful level matching and strict double
=20
blindness must still be maintained. I suspect that such a test migh=
t
=20
uncover differences that short term, instantaneous switching doesn'=
t
=20
reveal.
=20
=20
=20
As long as I've been reading RAHE (which is going on 15 years, I thin=
k), I've seen this belief expressed. One of these believers ought to try i=
t sometime. Perhaps they will teach the world of psychoacoustics something.=
(I am not holding my breath.)
=20
=20
=20
I have tried it.=20
=20
=20
=20
No, you haven't, not with enough rigor to pass the laugh test.
That's amusing. Anytime you strict "objectivists" hear of a result=20
that goes against your dogma, it's always "The test was not=20
conducted properly" Or "the test wasn't rigorous enough." I'll
say one thing. It certainly keeps you fellows safe and sound.
It's clear from your posts that you don't have a clear grasp of what is re=
quired to conduct a scientifically valid DBT.
Well, then Bob. When you get finished laughing at me, why don't enlighten u=
s=20
as to the correct methodology for conducting scientifically valid audio DBT=
s?=20
And if there are any differences, one has a much better=20
=20
chance of uncovering them if one really listens to the devices being=20
=20
auditioned. You can't do that when two devices are being swapped=20
=20
out for each other every few seconds (or even every couple of minutes).=
=20
=20
=20
=20
There's solid science that says the opposite is true--comparing brief sni=
ppets of sound and switching quickly between them actually makes the test m=
ore sensitive to differences, not less. Sean Olive's speaker preference tes=
ts for Harman use per-speaker presentations measured in seconds, not minute=
s. If longer presentations worked better, it would be in Harman's economic =
interest to use them.
Here is just another reason why I think that DBTs for audio are a flawed=20
methodology. How, for instance, are you going to tell whether
one DAC resolves image specificity better than another with short snippets?
What if the "snippets" are studio produced with multi-miked and multi-
channel performances that HAVE NO image specificity? How is switching
between two DACs, for instance every couple of seconds going to show=20
the imaging characteristics of the DUTs, even if that info is present in th=
e
program material? It can't. No wonder Scientifically valid DBTs so often
return a null result with the finding that everything sounds alike.
Thanks for confirming my doubts, Bob.=20
=20
=20
=20
As long as the auditions are truly double-blind, and the levels are car=
efully=20
=20
matched to less than a dB,
=20
=20
=20
And here is clear evidence that you don't know how to do a DBT. A 1 dB di=
fference is way too large. Try 0.1 dB. Also required is a forced-choice for=
mat and a statistically significant result over a meaningful number of tria=
ls.
=20
=20
=20
bob
Would you mind telling me EXACTLY how anybody is going to match=20
levels to 1/10th of a dB? Even the most expensive audio equipment=20
doesn't have potentiometers on them that have that kind of resolution,=20
because audio doesn't need that kind of resolution. It can be done to less
than one dB easily enough, maybe even 1/2 of a dB is doable, but 1/10th?
That would require aerospace-spec 10-turn pots, and they're not exactly=20
cheap, or all that easy to source. And some here say that DBTs are easy
to set-up. Gimme a break.
|