"Dick Pierce" wrote in message
et...
"Harry Lavo" wrote in message
news:mopPa.27198$OZ2.4772@rwcrnsc54...
Thought you guys might find the following interesting.

http://www.rogernichols.com/EQ/EQ_2000_02.html

I also suspect that responses will be predictable, but perhaps not.

As I mentioned in my previous response, Harry provided the first
"predictable response" to his own post, simply in what I would
assert is his clear mischaracterization of the topic as "Ears vs
Instruments."

In reading Mr. Nichols' text, we find, in fact, no such conflict
between ears an instruments. Why, because nowhere does he mention
any attempt to use the relevant measurements. I have no doubt that
Mr. Nichols' experienmce is quite real, and, from other sources,
I have no doubt of the problem in the stamper that could lead to
the problems. But Mr. Nichols simply failed to carry out any relevant
measurements. He talks about looking for gross errors and finding none.

He jumps to the conclusion, based on almost no objective data, that
the problem is jitter. He may well be right, but he has no confirming
evidence.

There would be plenty of ways to confirm his diagnosis: simply looking
at the noise floor would be one way, and actually (gasp! horrors! zut
alors!) actually MEASURING the jitter would, i might humbly suggest,
be yet another.

But that was NEVER DONE!

So where, Harry, is the supposed conflict between "ear" and
"instrument" that you see, when, in effect, no "instrument" was
used? He never said that appropriate instrument measurement failed
to reveal the problem. What he DIDN'T say was most eloqient: he never
made ANY relevant measurement.

A completely similar argument could be raised if he measured the
bejeebers out of it and never once listened to it. If jitter was the
problem, you'd see it trivially in a high-resolution spectral plot,
you'd see it trivially in a straight jitter measurement. Now, with
that in hand, where is it reasonable to title a post "Instrument vs.
Ear"?

Again, I am sure Mr. Nichols' experience is quite real. I am also
sure that the conclusion you seem to want people to infer is simply
unsupportable from his data, because he has NO data on "instruments."
Indeed, he does not state otherwise.

Where's the conflict?


The conflict, Dick, was that SONY and the production plants were all using
conventional measurements that they *thought* provided adequate quality
control to insure that the finished product would sound like the master.
And these measurements were all based on the "bits is bits" assumption.
But the ear/brain combo said "something doesn't sound right". And by
eliminating possibilities, the problem was narrowed down to the point where
the *important* variables creating the problem were eliminated because
somebody else had apparently determined the same thing and made sure those
variables were eliminated. Apparently that Denon plant and the JVC K2
people (that's their XRCD24 line, btw, I believe) trusted their ears too, at
least the JVC people claim to use rigorous listening as well as measurement
in setting up their system and Denon is routinely praised by Audiophiles for
their sound quality (and where John Eargle is (or was?) chief engineer).

So, could it be measured? I'm sure, if one knew what could cause the
problems in the first place and then track them down. But were the right
variables measured? Not routinely by the production engineers convinced
that "bits is bits" and if you don't measure bit errors "do we really have
to listen?" Yep, when all is said and done, there is no substitute....at
least until one has proven that *all* the audible variables are under
control.

p.s. the "they weren't dbt'd and probably the differences were imagined"
chorus has started. But I am pleased that your are focused on the measuremen
ts, because that was my own focus. As I am sure you will yourself agree,
sometimes their *are* large differences so apparent to a group of trained
people that the differences can be accepted as a given. In the antidotes
portrayed, there were tests and comparisons done which resulted in a "no
difference" when it would have been quite possible to have been biased in
favor of finding a difference...for that would allow the problem to be
solved. Instead, "no difference" was declared and the search continued. I
frankly am convinced that in this case the differences were real, and am
more interested in the QA measurement scenario.