[Arny Krueger[_4_]]

"Andrew Haley" wrote in message
...

Audio Empire wrote:

In a blind test, everyone involved could always pick-out the
HDTracks 24/96 version from the three CDs as sounding much better
than any of the CDs. Not particularly scientific, I agree, but it
was so blatantly obvious whenever the 24/96 file was played, that I
don't think it would have made any difference in the results had the
test been double-blind. I.E, the listeners couldn't tell which
version was playing through the speakers at any one time from any
criteria other than how the playback sounded. All three were level
matched to the resolution of the HP 427A rms audio voltmeter, so,
loudness wasn't the issue. What I don't know is why the computer
file sounded better ((lack of read errors, maybe?). There are simply
too many steps involved in transferring an analog master tape to
digital (especially in three separate CD releases done over almost
20 years) to say that the "hi-rez" file (transferred in 2010,
AFAICS) sounded best due to the fact that its 24/96.

It's easy enough to find out. Do a near-perfect downsample of that
24/96 file to 44.2 (easy enough these days in a digital audio editor),
resample it back up to 24/96, and see if anyone can tell the
difference. I'm sure PC/ABX would make it easy.

Not only is it easy, its been done by a number of folks, particularly
advocates for high resoluation recordings. The Hydrogen Audio Forum has been
a common venue for this sort of activity. Positive results are rare and when
they exist, they are generally traced to procedural errors.

One common procedural error is that people don't properly synchornize the
two recordings being compared. It needs to be done within a few
milliseconds. This can be tough, depending on the details of the experiment.
Another problem has been the use of an audio chain with excessive high
frequency nonlinear distortion. Another common problem is audible switching
transients that depend on the actual alternatives being switched between at
that instant.