. [snip]Eliminate witness subjectivity? You have a protocol that makes all witnesses have the same taste? Can it make them all want to marry the same woman, also?[snip].
I'm sorry Rod; in posting, one always assumes some basic understanding, but sometimes I cut too many corners.
Of course you can't make someone's subjectivity disappear.
What I meant was that the test protocoll will be designed to eliminate the influence of the listeners biases, prejudices, subjectivism etc from the results.
Actually you can take it one step further and design a controlled subjective test which also can give reliable and repeatable results. SY has written about that, I believe, in another thread.
jan didden
No. John has said that he cannot hear differences *unless* he can 'peek' and knows which is which. I can do that too. Anybody can do that, even with blocked ears
jan didden
Jan, I am sure you could systemetise the testing to eliminate bias and prejudice, HOWEVER, your results are still encoded by the ability of the subject to discern the difference between one equipment playing and another. This ability will vary from subject to subject, not to mention environmental conditions, and even emotional conditions.
Still, let's say you perform some tests and record the answers. Will the designer of the next BLOWTORCH use such data aid his work? No, I don't believe it at all. First, the data is invalid for any other circuit environment than the original test [as a true son of Apollo, you must accept that]. What use are tests based on stale circuits?
Second, will an outstanding professional designer, like John Curl, trust the results of test upon subjects he does not know and trust? No, he will trust his own perceptions, and combine them with engineering design talent, and a lifetime of professional knowledge gathering to produce an outstanding amplifier.
I am not casting doubt on your ability to produce tests, and interpret them. I am highlighting the fact that their practical usefulness in amplifier design, which is the central theme of our thread, is very small.
Still, let's say you perform some tests and record the answers. Will the designer of the next BLOWTORCH use such data aid his work? No, I don't believe it at all. First, the data is invalid for any other circuit environment than the original test [as a true son of Apollo, you must accept that]. What use are tests based on stale circuits?
Second, will an outstanding professional designer, like John Curl, trust the results of test upon subjects he does not know and trust? No, he will trust his own perceptions, and combine them with engineering design talent, and a lifetime of professional knowledge gathering to produce an outstanding amplifier.
I am not casting doubt on your ability to produce tests, and interpret them. I am highlighting the fact that their practical usefulness in amplifier design, which is the central theme of our thread, is very small.
Well, if you do a controlled test between, say, a BT and a SONY xyz, and the result is that the majority of listeners cannot hear a difference, I would think that's a pretty significant pointer for amp design.
Again, lots of people, when given the choice, would probably get a BT instead of the SONY. But with these test results in hand, we now know that the basis for that selection is not the audible performance. Isn't that interesting to know?
Again, anybody listens to what he likes, and selects and buys what he likes, etc. But if you say: this amp sound different from that one (which is the first requirement for different preference of course), that can be tested. Simple as that.
jan didden
No, that's a pretty significant pointer for the 'controlled' test methodology, and is valid for the one and only test conditions, including spatial acoustics etc. The pointer is that the hypermarket buyer would not need to go for the BT.
Even if ONE listener could reliably distinguish between a BT and a Sony, that would be useful to know. So far, nope, no-one has been able to demonstrate anything even vaguely like that. Not one person.
I'm trying to discover how many ways audiophiles including tinkerers can invent to argue about how many angels can dance on the head of a pin. This particular one is not a new one, it's been going on for well over a decade especially in regard to crossover networks. Nevertheless, so far I've tallied 137 of them. How about some originality for a change.
When communication equipement is tested, as well as all the electrical test, subjective listening tests are done. These employ such things as modified ryhme tests. To get some statistical evidence, in our case, all participents have a hearing test to determine their hearing ability and more important any frequency drop outs. Then there has to be quite a large number of participants, the more the better. There are other criteria to follow, but subjective testing can be done (certinally for communication gear) that gives results that are objective enough for even some of the most demanding customers. These are also suplemented by using some expensive test gear that imitates a human torso and how we pick up sounds (bone conduction etc) that gives further a reference.
Applying these tests to audio gear would be interesting and challenging, though costly, and possibly not always in a manufacturers (or audiophile theorists) favour.
Applying these tests to audio gear would be interesting and challenging, though costly, and possibly not always in a manufacturers (or audiophile theorists) favour.
I agree. So you'd set up the test in an 'average listening room', with agreed speakers, etc. I'm not saying this is perfect, but it is so much more reliable than an unknown, isolated listener, in an unknown environment, who has a stake in his equipment, and does a totally uncontrolled test whether his cap sounds better than another cap. Such is test is really worthless for anyone except the one tester.
If I do a well controlled test in an average listening room, with accepted good speakers, between a BT and a SONY xwz, and the result is that with 99% confidence you can say there is no audible difference, that's a significant result.
If it indeed is true that a BT is so much better, has such bigger soundstage, has so much more details than a SONY xyz, I would expect it to easily be differentiated in such a controlled test, like a 98% confidence that there IS an audible difference.
jan didden
Last edited:
Fully agree. You CAN do controlled, subjective tests that give usefull, repeatable results, in your case whether a customers pays for the gear or not.
There are some examples of such tests done by the audio industry that are documented; Floyd Toole did some speaker tests when he was still with Harman. One test result I remember was that the preference rating in a set of 4 different speakers depended on whether the participants knew which speaker was playing (sighted test) or not (blind). Very generally speaking, the larger the speaker the better it did sighted.
Interestingly, in this tests, some participants were experienced listeners that declared beforehand that they knew all about prejudice and bias and that they were convinced they could be 'objective' and judge the speakers on sound only. Turned out not.
jan didden
How do you or F. Toole assure adequate listening position for a large set of subjects during the speaker listening test? Usually you are able to do it for 2-3 people, no more, in case we speak about serious high end listening test with all the spatial and localization information provided. Without that, with group of 20 people sitting between the speakers, the test is just useless.
No you are right. Even if you have only one listener that reliably, with high confidence factor, can hear a difference in a well-controlled test, that's a clear indication that there IS an audible diference. QED.
If he/she can't, that doesn't mean there is no difference, only that we don't know yet whether there is one.
jan didden
Two stars should read "B and S"
For some silly reason, I can't edit my previous post.
I pasted the document name from the very document and the letters ** ended up as **. A least in my browser. And I can't edit that....
Edit: Aaahhh, the letters B and S together mean something special in English.....
So they end up as two stars. Really silly :-(
For some silly reason, I can't edit my previous post.
I pasted the document name from the very document and the letters ** ended up as **. A least in my browser. And I can't edit that....
Edit: Aaahhh, the letters B and S together mean something special in English.....
So they end up as two stars. Really silly :-(
Last edited:
I scanned that document 2quad, thanks "METHODS FOR THE SUBJECTIVE ASSESSMENT OF SMALL IMPAIRMENTS IN AUDIO SYSTEMS INCLUDING MULTICHANNEL SOUND SYSTEMS". I was interested in the test method as it was almost an exact description of what was carried out & described below.
It's interesting, when evidence of blind testing is put in front of the same people arguing for it on this thread, it is rejected as having been coached. There's no pleasing or satisfying some people
See here
I'll quote it
It's interesting, when evidence of blind testing is put in front of the same people arguing for it on this thread, it is rejected as having been coached. There's no pleasing or satisfying some people
See here
I'll quote it
Last edited:
Just so as you get the level of disingenuous approach that is being used here -
So I see the same avoidance tactics being used yet again here!
SY's post
SO SY as Rod has asked you already - Was there a a test between BT & Sony that you mentioned?
So I see the same avoidance tactics being used yet again here!
SY's post
doesn't exactly ring true now does it?
SO SY as Rod has asked you already - Was there a a test between BT & Sony that you mentioned?
Last edited:
It's beyond silly. But that's what happens when uptight, ultra-religious prudes are in control of the Nanny Filter.
se
- Status
- Not open for further replies.