Did Amir Change Your Mind About Anything?

@nonoise 

Imagine if this were a watch forum and someone went to extreme lengths showing you how your brand of watch is way overpriced for what it does and that a cheaper one is much more accurate and he had the graphs to prove it!

You have the start of a great analogy.  Sadly you got the middle and ending quite wrong.

Do imagine that this is the watch market except that all that anyone cares about is accuracy of time keeping.  Manufacturers learn this and claim that their watches are more accurate than any other watch.  Except, unlike the real watch market, they provide no measure of that accuracy.

Imagine further that companies realize that since no proof is needed, any all things can be sold under the guise of better accuracy. Companies come to market selling aftermarket watchbands that they say improves fidelity.  Ergo, they can charge more for some of these bands than you can buy entire watches.

This goes on for a while until a retired engineer, technologies and manager from said watch market says to people on his watch forum that he has highly instrumentation to measure such accuracy.  He starts to measure a few watches he has bought and shows how some of very accurate while the others are not even though they cost more money.

He publishes that result and next thing you know, watch owners want to know where their watches land.  So they start to send him their watches -- some cheap and some very expensive and he tests and publishes them.  Soon it becomes obvious that how much you paid for something does NOT at all predict how accurate said watch is.  And that the claims made by companies can trivially be shown to be wrong.

Watch owners love the clarity the above testing brings to market and increasingly support that activity by visiting the site, sending more product, and helping offset the cost of running this activity.

You would think every watch owner would be in favor of this.  But no, prior to this development, folks were looking at a watch and without any evidence, claiming that they have found the most accurate watch.  But here comes the above testing showing that to be the wrong statement.

A logical person would abandon the old ways and join the new.  They would not go on another forum and make up accusations that are trivially shown to be wrong.  For example, claim gets made that the engineer above doesn't even wear a watch.  All he does is look at the graphs of watches.  He shows that he not only has a watch, but multiple ones at all price points.  No matter. 

Folks start to get personal with him.  They accuse him for being in this thing for money.  They can't find any evidence of it but hey, if you make the accusation often enough, maybe it sticks.

In a direct one on one exchange, the very same folks don't have any facts to back their assertion of being able to tell how accurate a watch is based on just wearing said watch and measuring how long it takes for an apple to fall from the tree by counting under their breath.  No amount of telling them that is not accurate enough to count to fraction of a second gets them to listen.

So here we are.  We, I and literally tens of thousands of your audiophile friends try to bring more data and science/engineering to the table.  You don't like that?  No big deal.  Just don't make contrived analogies as if that will amount to anything.  

Ah, the video where you are sitting in front of all the electronic analyzers that you used to visually real time analyze the ABX signals?

There is no such video.  Stop making stuff up AJ.  I have never, ever used an analyzer when taking these blind tests. 

You don't even understand the nature of these tests and whether an analyzer can even help you.  Take the Archimago test.  That test relies on bit depth of content, not anything that you can analyzer with an audio measurement device.  If you don't believe me, go ahead and show the difference using said analyzer.

Why is it that you are not complaining about @kevn?  Did he or did he not pass the test of high-res vs standard not just by himself, but a few of his friends?

You are not going to answer that, are you? 

@soundfield 

Would you do a proctored one at PAF 2024? Or only doctored?

I would *love* for you to set up such a blind test for everyone who comes to your room there.  Are you going to do that and publish the results at the end?

Or is it that you don't want to alienate potential buyers of your speakers so that is the last thing you would want to do?

@soundfield 

And flat FR is sometimes boring.

Virtually useless and perhaps harmful for deciding which sounds better.

As a speaker maker, what do you have to say about this?  Is CEA-2034 frequency response measurements "virtually useless" and "perhaps harmful" in determining the fidelity of the speaker?  

Wise and right on the target...

This shows a complete misunderstanding as to the nature of double-blind testing in audio, such as ABX testing. Such tests are not designed to test the listener - that’s the role of an audiologist. The listener isn’t under test at all. What’s being tested is whether two signals can be distinguished under the conditions of the test. That’s why the best blind test programs include multiple listeners and multiple trials.

Some might argue that, if a specific listener claims to expect a difference between, say, a hi-res and lo-res signal, that an ABX test with him is "testing the listener." But that’s mistaken. Such a test could only reveal whether that listener could distinguish a difference under the conditions of the test. Again, this why is why multiple tests yield more useful information.

It’s rather odd that Amir is so preoccupied with conducting measurements that he sometimes doesn’t bother to listen to the devices he tests, and yet on the other hand issues such proclamations about the tests he’s claimed to have "passed."

As an aside, conducting a proper audio double-blind test is tricky business. I've seen it done and it's not as easy as it looks. When they’re well conducted, I’ve found that many differences become harder to distinguish than might be expected. When they are improperly conducted, such a test has no advantage over a sighted test and can yield misleading results.