Are audio companies all involved in a huge conspiracy? - Page 78

I take my leave. You can all discuss your DBTs now, no more Heinrich to disturb the peace. Don't worry about the interpretation of the evidence, just brush it aside and make declarations by fiat. That is what you people are good at, so have at it.

Thank you Chu Gai for being kind throughout this discussion. I'll rather read than participate as I don't feel I have anything more to add.

Farewell.

There is a god after all.

I take offense to that. Now I'm really going.

Farewell.

I

I run this Behringer EP4000 as well as a few other amps - Arcam AVR600 / Exposure 2010S2 / Cary SLI80sig. For my Magnepans, none of the other amps sound like the Behringer does - the EP4000 is an excellent amp. I really don't know what to believe when I read specs but from my home tests, the Behringer is clean, strong and kinda charming to look at. I did the fan mod which cut the noise by 75%. Overall, one of the great choices in my audio history.

I think you should be skeptical of claims that such phase cancellation is audible. You don't hear it because you're a human, and our ears are not very sensitive to phase differences. This is like the people who are absolutely convinced that center speakers with several side-by-side drivers have got to sound bad because of those cancellation nulls. They're sure they hear them because instruments and physics say they're there. How could anyone not hear them?

Is that anything like being sure that two amplifiers must sound different because you're convinced they must?

Yes, if you were convinced by instrumental tests. Perceptual claims require perceptual evidence.

You basically ask us to convince you and we tried for over last 20 pages but you are not convinced. We even suggested that you try it yourself if you are that unconvinced but you won't try it. It's clear that you jumped into this thread with your mind already made up based on your opinion (opinion based on opinion, what a concept!). You've been trying to denigrate the existing DBT results because you personally don't like the result. Sorry to tell you that it didn't work and it hasn't worked in the past (you aren't the first one to try it).

Please feel free to share if you know of a better way to remove bias when doing listening test of couple audio gears.

It's not about a sensitivity to phase differences, but that as you stand up there will be different path lengths from each driver to ear which means that at certain frequencies you will have a notch due to the signal from each driver being out of phase at that position. This will cause a notch, or amplitude variation. However we're not all that sensitive to those either, but certain selections and types of music will show it up.

Not familiar with the codec test as I wasn't very concerned about the new HD video alternatives at that time. Anyway, results such as this aren't surprising to me when testing for preference in the manner you describe above. Had the test been set up as a "detection of difference" test where the options were binary, either "same" or "different" with random master-self controls, variances in preference would have fallen out and more biases removed.

But you're certainly right that biases remain even in controlled testing. One of the more important ones which has been mentioned is the possibility that objectivists will vote "same" because they expect the DUT's to be the same. Throwing in some "known audibly different" DUT's as controls for that is possible too, however, and at least you'd know not to count such data if the listener can't reliably distinguish devices known to be audibly different. In fact, I think that would be a nice inclusion in any such DBT.
I suppose that isn't too surprising in a test for preference. After all, the initial assumption (not tested for, though you might be able to tease such out of the data) was that all the codecs were visibly different. We hope humans prefer accuracy to the original, but as has been touched on here, that isn't always the case as proven by people preferring all manner of odd-behaving tube gear, bizarre speakers, and perhaps even my preference of surround extraction over "the way it should be heard" two channel.

Perhaps some codec's applied a sort of anti-aliasing smoothing effect that some found preferable to a higher frequency (noisier) information in the original? Who knows. There's no accounting for taste!

Reminds me of the old joke: Elvis himself entered an Elvis look-alike contest at a local restaurant shortly before his death, and came in third place.

The flaw with this little anecdote is that the facts seem to relate to a DBT that was not ABX, and was probably some form of ABC/hr or ABCD/hr, MUSHRA etc.

The alleged flaw with the cited test that was found cannot possibly happen in ABX tests.

http://en.wikipedia.org/wiki/Codec_listening_test

"
ABX test

In an ABX test, the listener has to identify an unknown sample X as being A or B, with A (usually the original) and B (usually the encoded version) available for reference. The outcome of a test must be statistically significant. This setup ensures that the listener is not biased by his/her expectations, and that the outcome is not likely to be the result of chance. If sample X cannot be determined reliably with a low p-value in a predetermined number of trials, then the null hypothesis cannot be rejected and it cannot be proved that there is a perceptible difference between samples A and B. This usually indicates that the encoded version will actually be transparent to the listener.

ABC/HR test
In an ABC/HR test, C is the original which is always available for reference. A and B are the original and the encoded version in randomized order. The listener must first distinguish the encoded version from the original (which is the Hidden Reference that the "HR" in ABC/HR stands for), prior to assigning a score as a subjective judgment of the quality. Different encoded versions can be compared against each other using these scores.

]MUSHRA
In MUSHRA (MUltiple Stimuli with Hidden Reference and Anchor), the listener is presented with the reference (labeled as such), a certain number of test samples, a hidden version of the reference and one or more anchors. The purpose of the anchor(s) is to make the scale be closer to an "absolute scale", making sure that minor artifacts are not rated as having very bad quality.
"

The Wikipedia article seems to be sufficient to make the differences in terms of purpose and execution among these three kinds of testing clear enough to the average reader. Somehow Amir missed the boat.

 

Amirm knows I don't read his posts so why he replies to them is a mystery to me, unless it is because he knows I don't see his replies ;)

 

As usual, he takes things out of context or deliberately conflates things, and this is another example. He not only says that what I said is untrue of ABX tests, but he then goes on to reveal that he doesn't know what an ABX test is!

I suspect that Amir does indeed know what an ABX test is, but is unable to restrain himself from the attempted conflation/deception. The element of the ABX test that avoids the problem he points out is the fact that an ABX test is a black-versus-white forced-choice test. This could be a problem if it were not true that reality is full of identical kinds of choices.

 

 

Yes, a good observation. 

 

Quote:

 

He often seems to snipe at my posts, saying I am wrong when I am not wrong and then 'proving' it by conflating what I said with whatever suits his agenda that day. I am more than happy to admit if I am wrong - I am basically here to learn - but it is annoying to be 'corrected' in this way when I am not wrong.  ABX: black-versus-white choice as you say, or as I said, two outcomes - the testee hears a difference or the testee does not hear a difference.

 

I'm glad you chose to reply to Amir, otherwise I wouldn't be aware of his post.

If you'd put down the hatchets from whatever thing that's gone on between you in the past, I can't see how you could possibly have any objections to:

If you're not the person doing the test, you cannot know his/her motivation. Might not want the test to succeed, may be too unfocused or uninterested to do their best and several other things.
It does happen that very strong significant results do happen by random, even if we statistically want to close out eyes to that when we get good numbers.

I had a german-test in highschool that had a very difficult passage (20 or more questions) with six alternatives. One guy that never cared to learn any german, just sitting off the time and taking the lowest grade wrote a-f on one of his pencils and rolled out his results. He scored full on that passage... our teacher didn't understand why the whole class were laughing their asses of when she complimented him on being the only one getting that part correct.

So, I do think you're a little too quick to fault Amir - he does have a point. Whether the point has any relevance is another matter, but he pointed that out himself too.

 

I said that in an ABX test there were only two possible outcomes: the testee hears a difference or the testee does not hear a difference. He said this was "untrue". It is not.

 

Everything he said after that was in reference to tests that were not ABX tests, thus irrelevant to the discussion. I am not 'being quick to fault him' - I am pointing out that what he said about what I posted was wrong.

 

 

I can take an ABX test where I clearly hear a difference but choose to score in such an average fashion that it seems like random guesses. Then the test does not tell whether I heard or not, it only tells what I voted.

"Jeeze, look, this one guy didn't notice the differences that everyone else could hear."

 

And what would the point of that be?

 

Not to mention you'd be the first to hear these differences in an ABX amp test of course ;)

In this thread - just to show you that your logic wasn't bulletproof. You only know what was voted - not why. You can phrase yourself in another way that doesn't open up for such issues. That's all.

Give me a Jadis SE300B vs Bryston 14B-SST and a speaker setup that's capable down to 10Hz and then I want infrasonic bass rich material and plenty of volume... Put in a glass of water and you could have photographic evidence of the difference.

Are both amps specced the same down to 10 Hertz?

I don't care. If you're grasping for some definition standard of what constitutes an ABX-test, then I really don't care. If for me relevant differences can't be tested, then the test isn't good enough.

Jadis frequency response looks like this:

I think the premise was to compare amps that are competently designed with certain specs and driven within the margins of their parameters.
The Jadis does not fulfill that demand.
We should be comparing amps, not tone controls.

You don't care. Well, that's nice, but it also violates the (often unstated because it takes too long) core requirements and usual caveats which qualify the statements often parodied as "all amps sound the same". If that undulating curve in the diagram you posted is the FR of the Jadis, there's no reason to expect a device with such a horrifyingly bad response curve to sound like anything else.

That kind of test has no relevance for real life. If you are limiting the test to what's possible to compare in that way, you ignore how things are in real life. And then you end up with a statement that "all amps sound the same" which only is correct within that small subset. While the truth is that almost every amplifier made can be detected when compared to the signal it was told to amplify and it's quite hard to make different designs color exactly the same.