Archived from groups: rec.audio.tech,rec.audio.opinion (
More info?)
Mr Le Gal (Google message 86, Aug 30) quotes Greenhill's final comments
on his cable test as a rejoinder to my text in my reply to Mr. NYOB:
"P.S. To prevent you from quoting phony references again here is one
for you to digest: (L. Greenhill, Monster vs Radio Shack:same gauge
cable, ABX/DBT comparison Stereo Review '83)
> Three out of 15 panelists scored correctly well over 50% and one had 81% positive result. Which proves that a few can surmount even the ABX obstacle race.
So much for "anyone,ever" (Mr.NYOB said that no one ever heard
difference between cables under ABX)
For Greenhill's comments refer to mr. Le Gal's message.
So what else is knew Mr Le Gal? Greenhill, a good 'objectivist"
that he was provided a nice, objectivist comment to suit the nice,
objectivist mag. "The Stereo Review". Indeed the *majority*of
his panel had 50% or less corrects- under ABX/DBT it all sounded the
same to them. Just as happened in all the other trials of amps,
preamps, cdplayers and dacs up to and including a very, properly
designed loudspeaker trial by Sean Olive (JAES,vol.51, No.9, p.806).
You ignored however the interesting part
Greenhill found one consistently accurate panellist scoring 81%, in 5
out of 6 trials, of 15 tests ech, called him the "golden ear" and
observed: "Obviously certain listeners whether through talent,
training or experience can hear small differences between components.
But the majority_ etc" He had two others who came very close to that
high score but said nothing about it. Instead, like all the other
proctors in similar trials, he created through a "mix them all
together" statistical sleight of hand a fictional Mr Average, who did
not hear much.
The fact though was that SOME could overcome the handicap of the DBT
protocol and did well. Better than I would have done because every time
I tried DBTiing with an ABX model I found that after four trials I no
longer knew if it was Rimski Korsakoff or his cockerel that composed
the snippet. But even if only one panelist hears a difference with
statistically significant consistency then the difference is out there,
real to him. That it may not be audible to a thousand others is not of
the slightest relevance to an individual making his high-end choices.
A virtuoso doesn't care if anyone else hears the difference between
his Strad and a music store violin. (I wonder if he'd pass an ABX or
if one of our "scientists" could provide measured specs. for the two?)
In his conclusions Greenhill did not comment about this
contradiction between his results and his "golden ear" comments.
One year ago in the RAHE he was invited by his editor Mr. Atkinson to
elucidate but he chose discreet silence.
I can already hear the parrot cry (I do not mean you Mr. Le Gal):
"I do not like this result. I want a repeat, and then a repeat again
and again till Mr. Golden Ear gives in and signs up to my revealed
faith."
Funnily enough the same people
are perfectly happy with Greenhill's very scrupulous statistical
protocol- as long as it gives them the results they desire and wish
for.
Ludovic Mirabel