Well the 3rd amp tells you something useful: if A1, A2, B are all scored differently, you need a new set of listeners. 🙂
Actually the guy who handed the DUTs over did not know which variant belonged to which marking letter.
What you call a tertiary control would be in fact a different test protocol (sort the odd one out) and would, given the constraints of this test, not offer an advantage.
Listeners were carefully selected by me due to their listening preferences and to succeed all five had to prefer the same circuit variant to be significant at 0.05 level (actual P(X=5 l H0) = 0.0313 .
Last edited:
No, it's 2 tests built into one. It provides a natural negative control, which is highly advantageous for reducing the effects of reviewer variability. If the reviewers are blinded to the fact that there is an odd one out, at least.
Yes, double blinding the labels is preferred, if you have the human controls in place.
Yes, double blinding the labels is preferred, if you have the human controls in place.
Last edited:
Daniel,
I think those who don't want the third unit, a matched unit to one of the others are those who really want to remove the controls from an experiment, that leaves it easy to throw in opinion and not have to have proof of results. It keeps this game going forever with no real conclusions, the endless arguments never end.
I did learn this early when working in my fathers medical lab, we always ran two sets of controls, one before and one after the patients chemical testing. If the two sets of controls did not match throw out the results and do them over, no if's, ands, or buts about it, test results were confirmed or not.
I think those who don't want the third unit, a matched unit to one of the others are those who really want to remove the controls from an experiment, that leaves it easy to throw in opinion and not have to have proof of results. It keeps this game going forever with no real conclusions, the endless arguments never end.
I did learn this early when working in my fathers medical lab, we always ran two sets of controls, one before and one after the patients chemical testing. If the two sets of controls did not match throw out the results and do them over, no if's, ands, or buts about it, test results were confirmed or not.
Yeah, Steven, (or is it Stephen, sorry can't remember!) it'd be ideal to have a positive control as well to get an even better characterization of the test.
I don't know if I'd call it malicious, but largely poor understanding of experimental design and latching onto ideologies hard (something we're all guilty of, so that's why we need the controls/good data to keep us in check).
I don't know if I'd call it malicious, but largely poor understanding of experimental design and latching onto ideologies hard (something we're all guilty of, so that's why we need the controls/good data to keep us in check).
@ DPH,
did you really write "No" ? 🙂
What you proposed is known as "triangle test" and it _is_ a different test than an "A/B- test" (known as paired comparison, abbreviated 2-AFC if forced choice or 2-AC if the no preference/difference option is included).
As said before, it exists a plethora of literature about sensory testing, obviously just to be ignored. 😉
did you really write "No" ? 🙂
What you proposed is known as "triangle test" and it _is_ a different test than an "A/B- test" (known as paired comparison, abbreviated 2-AFC if forced choice or 2-AC if the no preference/difference option is included).
As said before, it exists a plethora of literature about sensory testing, obviously just to be ignored. 😉
Last edited:
I disagreed that it was an odd-man-out test, as, perhaps to my misunderstanding, the reviewers *need* to know that 2 are in fact the same beforehand in an odd-man-out test? That changes the game entirely.
Whatever we choose to call these, it should still come down to doing good experimental design and controls. An A1, A2, B test is certainly different from an A, B test, but provides a lot more value and stats.
Whatever we choose to call these, it should still come down to doing good experimental design and controls. An A1, A2, B test is certainly different from an A, B test, but provides a lot more value and stats.
I will show you a case example, why "hi-fi" is not always that simple... (and that THD is insufficient)...
The quality of an amp (and the matching speaker) cannot bee seen only from one variable (a.k.a. THD). Referring to the amplifier listening test by XRK, I found that I don't like FH9 (it's a hexfet output). This amp and the quasi hexfet-bjt are outliers. Their performance are so different with the other 3 or 4 amps (which are close to each other)...
But FH9 has a great impact/dynamics. So it depends on what you want to hear. If you hear cymbals, the FH9 is perfect, because cymbals should sound like metal with big energy/dynamics, and most tweeter cannot do it well, so with its known character the FH9 is more hi-fi in reproducing the cymbal... 😱
But, man, there are many others in music beside instrument's sound...
Exactly. This is why some of us smile when someone claims to have "measured" something by ear. However, if two systems are each good enough to sound like the original sound then they should sound like each other too. Part of the reason to have short time lags is that this doesn't give the ears time to recalibrate their 'white balance'.vacuphile said:
Did anyone say that THD is sufficient?Jay said:
An amp which reproduces cymbals but not much else is not hi-fi. When I mentioned using a violin I did not mean just a violin, as an amp which can only reproduce violins would not be hi-fi either. However, being able to reproduce something is better than reproducing nothing.
OK. So peeking at two same looking boxes with DUT in them is Ok so long as we do not know the identity of what is inside the box. But it does help the memory to be able to say A sounds different than B here and there etc. Long listening sessions, it is particulary needed to recall which you are hearing and can attach a character to that one (box A or B).
Anyway, where is Bob Cordell? I want his input on an article I found in my closet ---
Electronics World June 1997 by William de Bruyn... pg 476-478. Error Feedback in Audio Power
View attachment Cordell Diff.pdf
THx-RNMarsh
Anyway, where is Bob Cordell? I want his input on an article I found in my closet ---
Electronics World June 1997 by William de Bruyn... pg 476-478. Error Feedback in Audio Power
View attachment Cordell Diff.pdf
THx-RNMarsh
Last edited:
Curious now...
Did your preamp sound the same as his preamp on his system, from the lathe?? (a master?)
And, did the commercial pressing of the LP sound the same as the master using your preamp or his preamp?
No preconceived idea or trick questions, just would like to know.
_-_-
I joked with Olive once about how musical a particular component was. It made everything sound like a cello.
Yes. Not you. I wasn't disagreeing with you.
That's what I mean... Problem is, there are so many variables (if you can hear it and know it exists), so that amps may excel at different things, so that there is no hi-fi, only preference. This is what many people don't understand.
Now, on trusting ones ear, I have reasonable confidence in my ability to pick up on a bouquet of audible flaws and identify them by ear.
Like tweeters with gummy ferro fluids, cross over distortion, Kaiser Wilhelms, fr deviations, resonances, room modes, it will be a long list before it is complete. This comes naturally with many hours of listening, measuring, improving and debugging speakers with associated electronics. Others will have similar experiences. Ears are good for that and training can make them even better.
Flaws.
Like tweeters with gummy ferro fluids, cross over distortion, Kaiser Wilhelms, fr deviations, resonances, room modes, it will be a long list before it is complete. This comes naturally with many hours of listening, measuring, improving and debugging speakers with associated electronics. Others will have similar experiences. Ears are good for that and training can make them even better.
Flaws.
Last edited:
destroyer,
Just looking down the page at the adverts is enough to make you question our sanity.
Steven
Just looking down the page at the adverts is enough to make you question our sanity.
Steven
- Status
- Not open for further replies.