John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
I have heard a lot of opinion on this and several similar threads. However, it does not follow what I have found in my personal and professional life experience.
First, violinists really can hear the differences in violin quality. Mostly, the old instruments do better, and because they have collector's value, like an old painting, they are VERY expensive and hold their value, like a Marantz 9 tube amp or a Vendetta SCP-2D.
I do believe that is what I heard when I heard a particular Strad played by a master violinist, back in 1975.
However, I had heard other Strads before at the institute where I worked, for example, Zino Francescatti, who was my wife's former master teacher, and we often had dinner together as well as to hear him play, yet while I appreciated his violin sound, I was NOT knocked over. Yet, when one year later, this relatively unknown (to me) violinist played his instrument, I was knocked over!
I attribute this to my innate capability to discern audio differences, developed in selecting classical guitars to own and play, in my past.
I would say that many here, might do better than me in tasting differences in and truly appreciating fine wine. I KNOW that many here put their extra money into that area.
I don't, BUT I also remember buying a bottle of fine wine, at 'The Grand Hotel' in Montreux, where we lived at the time, and the wine steward, being friendly, slipped us a misplaced bottle that had at least 6 times more value than we were normally paying for, and again, I was knocked over! So much that I took the bottle home and saved the label. Yet, I am and never was a true wine advocate, like some here. Still, I had my wine tasting experience, and will never forget it.
It is the same when I sample audio systems brought to Las Vegas, once a year. Sometimes, I am knocked over, sometimes, I am not. Yet, I don't need, once again, that I have to do a blind test to tell the difference.
There is a forceful movement to tell us all that everything essentially sounds the same, but it is not true, BUT we can be put into a double blind testing situation, that will give us NULL results. IF this taught us a lesson, then why doesn't everything sound the same, AFTER the test is over? But the differences reappear, and we are back to finding out WHY?
 
Last edited:
Hi,

There is no such movement

If there is no such movement, what are Krueger, Noisaine, Clark, Lipshitz et al constantly writing and whinging about?

Like for example in this article:

http://www.nousaine.com/pdfs/To%20Tweak%20or%20Not.pdf

Which incidentally is a masterpiece in demonstrating how to conduct scientific tests, including positive and negative controls, serious statistical evaluation of the results (e.g. confidence interval) and arranging tests in a manner that minimise "test stress" - Not.

Ciao T
 
If there is no such movement, what are Krueger, Noisaine, Clark, Lipshitz et al constantly writing and whinging about?

I guess I missed the part about everything sounding virtually the same. Could you point out where any of the folks you name have actually said that?

In the popular magazine article you linked to, there is some detail on how the test accommodated listener preferences (the soi disant "test stress"). You may want to reread it.
 
Hi,

I guess I missed the part about everything sounding virtually the same. Could you point out where any of the folks you name have actually said that?

While those who like split hairs will argue that stating: "failed to demonstrate a difference in hundreds of tests" is not the same as claiming "everything sounds the same", it certainly is close enough for government work, McCarthyist Witch hunts and book burning...

I actually pointed JJ (a well known audiophile/audio researcher) to this particular piece of "journalism and scientific research" that also happened to run in Audio where I read it (if memory serves) in a long ago debate over at the Audio Asylum (of which you may have also been party) as I took exception to his defence of ABX testing.

AFTER he had read this particular piece of "journalism" and "scientific reseach" JJ also decided to take exception to his defence of ABX testing. I suspect he also spit out his coffee, however, that does not form part of the public record...

Now we may debate the precise semantics till the Cows and Bulls come home from whatever they are doing when they are out and about, but the intention of the torrent of printed paper that is spewed forth by said sources is very clear.

Another little example is here:

The Great Debate...and Then Some | Stereophile.com

I rather rest my case. Q.E.D.

Ciao T
 
I support John's position: no matter how many ordinary people you employ for double blind tests, doctor will better tell what is wrong with person's health. Especially good experienced doctor, like John Curl is.

Edit: however, sometimes couch of athletes can tell better than a doctor what is right, and can be improved, though... ;)
 
Last edited:
AX tech editor
Joined 2002
Paid Member
[snip]There is a forceful movement to tell us all that everything essentially sounds the same, but it is not true, BUT we can be put into a double blind testing situation, that will give us NULL results. IF this taught us a lesson, then why doesn't everything sound the same, AFTER the test is over? But the differences reappear, and we are back to finding out WHY?

No John, this is incorrect on several fronts.

There IS NO movement that says everything sounds the same.
Because everything most certainly doesn't sound the same.
I thought we agreed on that.

But yes, the audible differences heard in uncontrolled test sometimes disappear in controlled tests. And then they sometimes re-appear again in a sighted test.
Now, are you really asking WHY that is? Because we know this already for several decades.

jan didden
 
IOW, they never said any such thing.

But the thread was getting boring, logic is rational and uninteresting, so to proceed irrationality is more fun.

I suspect you all know the proofs that all odd numbers are prime. So here we have the mathematicians proof. "One is prime, three is prime, so by induction all odd numbers are prime!"

Or is this the other version? "One is prime, three is prime, five is prime, seven is prime. How many of these need to be done to convince you?"

Or the physicist's version, "One is prime, three is prime, five is prime, seven is prime, nine is experimental error, eleven is prime, thirteen is prime, additional research is redundant."

But many here seem to use the version, "One is prime, three is prime, five is prime, seven is prime, nine is prime, eleven is prime, thirteen is prime, therefore all odd numbers are prime."
 
Hi,

I suspect you all know the proofs that all odd numbers are prime.

....

But many here seem to use the version, "One is prime, three is prime, five is prime, seven is prime, nine is prime, eleven is prime, thirteen is prime, therefore all odd numbers are prime."

Back when I learned math 9 was not a prime, BWTFDIK.

But I can add another version to this proof:

"Nousaine/Krueger/Clark/Lipshitz et all have many double blind tests that show that in a double blind ABX test even serious mathematicians cannot tell 9, 15 or 21 from other primes, so all odd numbers are primes."

Ciao T
 
Status
Not open for further replies.