or Connect
AVS › AVS Forum › Audio › CD Players & Dedicated Music Transports › CD vs. Vinyl...the DeathMatch!
New Posts  All Forums:Forum Nav:

CD vs. Vinyl...the DeathMatch! - Page 3  

post #61 of 2578
Quote:
Originally Posted by classic77 View Post

...But if listeners can't accurately pick X from A and B, this does not prove that A = B.


You are absolutely correct. You can measure A and B and they will not be equal. Cut 1" off a cable and it will measure differently but is that small length change is audible? At what point is something audible? How about components, are they audibly different? That is what all the discussion is about since the 'golden ears' use audibility as the criteria for choosing components, primary criteria.

Interestingly, people do perceive differences, in a DBT, when A=A, nothing is changed, at a pretty high level. I wonder why that is?
post #62 of 2578
Quote:
Originally Posted by classic77 View Post

From: http://www.hydrogenaudio.org/forums/...howtopic=16295

Rule 1 : It is impossible to prove that something doesn't exists. The burden of the proof is on the side of the one pretending that a difference can be heard.

What this is quite obviously saying is that it's up to the pretender to prove that the difference exists. But if listeners can't accurately pick X from A and B, this does not prove that A = B.

Um, all it proves is that they are indistinguishable from each other to the human ear.
post #63 of 2578
Quote:
Originally Posted by classic77 View Post

I'll tell anyone that our brains are not 100% reliable at remembering sound quality, especially as time passes.


I think you missed what I was referring to why I told you to try to tell that to the NRC. I was referring to the validity of DBT that you had issues with, not that memory is short or long. We agree.

But then, what happens to memory in a sighted comparison? It is still just as short, yet, listeners have no qualms with finding differences over long time passage.
post #64 of 2578
Quote:
Originally Posted by m. zillch View Post

Um, all it proves is that they are indistinguishable from each other to the human ear.

Nope, failure of the test proves nothing.
post #65 of 2578
Quote:
Originally Posted by classic77 View Post

Nope, failure of the test proves nothing.

OK, to be more specific, it proves that using that test track, on that listener, through that stereo system on that set of test trials the difference was indistinguishable.

Or are you saying that when people say, "I can't tell any difference." they may be telling a lie? Please explain.
post #66 of 2578
Quote:
Originally Posted by CharlesJ View Post

You are absolutely correct. You can measure A and B and they will not be equal. Cut 1" off a cable and it will measure differently but is that small length change is audible? At what point is something audible? How about components, are they audibly different? That is what all the discussion is about since the 'golden ears' use audibility as the criteria for choosing components, primary criteria.

Yes but in a DBT, measurements are irrelevant, that is if one can can accurately pick X from A and B. Measurements can be used to form a hypothesis. How small a change of components is audible? So many variables, so very hard if not impossible to determine.

Quote:
Originally Posted by CharlesJ View Post

Interestingly, people do perceive differences, in a DBT, when A=A, nothing is changed, at a pretty high level. I wonder why that is?

I'm not sure what you mean here sorry.


Anyway here's an interesting article where a "Blind Test" (I assume it was a double blind test) got bi-wiring accepted by a speaker designer. I found this as I own a couple of his speakers. http://www.legendspeakers.com.au/inf.../biwiring.html
post #67 of 2578
Quote:
Originally Posted by m. zillch View Post

OK, to be more specific, it proves that using that test track, on that listener, through that stereo system on that set of test trials the difference was indistinguishable.

Or are you saying that when people say, "I can't tell any difference." they may be telling a lie? Please explain.

By definition of a DBT, and this is from the resources I have examined, only the passing of the DBT can be accepted as proving something. Failure does not prove anything. If listeners can't accurately pick X from A and B, this does not prove or imply that A = B.
post #68 of 2578
Also remember in a DBT a listener must choose! Even if they have no idea, they can't say I can't tell the difference, they must answer A or B.
post #69 of 2578
Quote:
Originally Posted by classic77 View Post

If listeners can't accurately pick X from A and B, this does not prove or imply that A = B.

You are losing me my friend, how does your use of the word "pick" differ from my use of the word "distinguish". To me they are synonymous. Please explain.

To me you are writing,"Just because people can't hear a difference between X, A, or B doesn't mean that they all sound the same".... this makes sense to you?
post #70 of 2578
Quote:
Originally Posted by m. zillch View Post

You are losing me my friend, how does your use of the word "pick" differ from my use of the word "f listeners can't accurately pick X from A and B, this does not prove that A = B.". To me they are synonymous. Please explain.

To me you are writing,"Just because people can't hear a difference between X, A, or B doesn't mean that they all sound the same".... this makes sense to you?

A DBT is not a "can I hear a difference or not" kind of test. You must choose X from A and B.

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

Now for example I could compare CD to Vinyl. But guess what, I could be biased towards them being equal. So much in fact that even if I can pick X from A and B, I'm going to lie and simply choose A,B,A,B,A,B,A,B....etc. This way the hypothesis that CD is better than Vinyl is sure to fail, and you'll roughly get 50-50 results (like tossing a coin). So how can one say that this proves that CD and Vinyl are absolutely no different, equal, when I might have just lied my ass off? You can't.
post #71 of 2578
Classic77.

Don't get confused. DBT are extensively used to make millionaire decisions in food industry, medicine and some other areas where sensory tests are mandatory. They are actually conclusive!

I think Mr Hirvonen's Master Thesis would a be a great help for you to understand ABX tests and subjective audio test methodology.

http://www.acoustics.hut.fi/publicat...rvonen_mst.pdf

Enjoy!
post #72 of 2578
Crawford provided no details so it's impossible to replicate. It wouldn't surprise me if his prototypes were rigged though and besides, since when has Linn taken such a high road? For the full discussion, click on... http://groups.google.com/group/aus.h...c3c4cf9885c984

When a person wears multiple hats and one of them is pitch-man for the company, we should readily question the motives. After all, smoking doesn't cause cancer and lung ailments, does it? I seem to recall all the heads of the tobacco companies saying so before a congressional committee or two.

Quote:


Failure does not prove anything. If listeners can't accurately pick X from A and B, this does not prove or imply that A = B.

That's true. It's also true for dowsing, faith healing, and the like. A person's adamancy to hold on to positions is often proportional to their buying into something.

Quote:


Also remember in a DBT a listener must choose! Even if they have no idea, they can't say I can't tell the difference, they must answer A or B.

So modify it and select the 'No Friggin Idea' button. You still need to run a sufficient amount of trials to achieve statistical signficance. OTOH, you can run what's called a triangle test, where 2 are the same and one is different. All the time. No tricks. Here you make a forced choice as to which is different. If it's strictly due to chance, your success rate is 33%. Done all the time in the food and beverage industry as well as other areas.

Quote:


So how can one say that this proves that CD and Vinyl are absolutely no different, equal, when I might have just lied my ass off? You can't.

We rely upon your integrity.
post #73 of 2578
I also found the following in a USENET posting by Crawford.

Quote:
My "data" is personal - hence I deliberately used the phrase "in my
experience" - and you can take it or leave it as you wish. It begins when I
first adopted biwiring at Linn on the Nexus loudspeaker in 1986 and had to
get it past the listening panel dominated by Linn's sales guys. They had
been rubbishing biwiring publicly because of the NIH (not invented here)
syndrome and, being as self-opinionated as Ivor, were not easy to convince
they might be wrong. However, one brief double blind listening test (where
all we did was to change the biwiring links, leaving 2 sets of leads
connected to the crossover) convinced them they had to publicly eat their
words (not easy for Linn to do and something for which they never forgave
me)! And all my experiences since then have been similar eg a couple of
months ago when we first played Legend's new Kama 2 loudspeakers with and
without biwiring at Audio One in Sydney, Julian Eade spontaneously commented
that it should convince even the most prejudiced doubters of the benefits of
biwiring.

Note, that he says there was only ONE test. Now, really!
post #74 of 2578
Quote:
Originally Posted by classic77 View Post

But guess what, I could be biased towards them being equal. So much in fact that even if I can pick X from A and B, I'm going to lie and simply choose A,B,A,B,A,B,A,B....etc. This way the hypothesis that CD is better than Vinyl is sure to fail, and you'll roughly get 50-50 results (like tossing a coin). So how can one say that this proves that CD and Vinyl are absolutely no different, equal, when I might have just lied my ass off? You can't.

Oh, I get you now. Like I asked you in my post yesterday, "Or are you saying that when people say, "I can't tell any difference." they may be telling a lie? Please explain." You are concerned that people may cheat and lie. Instead of really trying to tell any difference they'll just fill in their score card randomly. Well why didn't you just say so back then?

Yes, you are correct, people who lie or cheat during a test make the results worthless. But I'll tell you one way that they can not lie: They can't claim proudly that because they have excellent listening abilities, that they can identify whether X is actually A or B (statistically), when in truth they are unable to demonstrate this ability through A/B/X testing. That kind of lie will be caught by the very test itself. Don't you agree?

Like you quoted,"The burden of the proof is on the side of the one pretending that a difference can be heard."
post #75 of 2578
Quote:
Originally Posted by m. zillch View Post

But I'll tell you one way that they can not lie: They can't claim proudly that because they have excellent listening abilities, that they can identify whether X is actually A or B (statistically), when in truth they are unable to demonstrate this ability through A/B/X testing. That kind of lie will be caught by the very test itself. Don't you agree?

Sounds useful. There is certainly plenty of that kind of lying going on around here.
post #76 of 2578
Quote:
Originally Posted by PULLIAMM View Post

Sounds useful. There is certainly plenty of that kind of lying going on around here.

Why on earth do you think that it is necessary to inject that ridiculous and useless comment into their discussion.

Whatsmore, you make your opinions very clear, and they are concrete, although not well reasoned. They are also clearly coming form someone with very little extended experience with any higher end equipment and the intricacies therein.

If I were someone that just a few days ago was called to task with a bet the way you were when you completely backed away, I would be ashamed and embarassed to go on this way.
post #77 of 2578
Quote:
Originally Posted by Harrypt View Post

I would be ashamed and embarassed to go on this way.

Then why do you go on that way?
I think it is rather revealing that you are acting all defensive when nobody mentioned your name.
post #78 of 2578
Quote:
Originally Posted by classic77 View Post

Anyway here's an interesting article where a "Blind Test" (I assume it was a double blind test) got bi-wiring accepted by a speaker designer. I found this as I own a couple of his speakers. http://www.legendspeakers.com.au/inf.../biwiring.html

I wasn't at the test, but I can tell you that 99% of people testing for the audibility of bi-wiring make the exact same fundamental mistake in their test methodology which "can" and does at times bias the results. It doesn't matter if the test is "blind" or not as you will soon see. By a show of hands, how many would attempt it in the following way? [We're not talking about blindness etc, just how does A differ from B]:

A= single set of speaker wire from amp to bi-wirable speaker with its gold jumper straps bridging high +/- terminals to low +/- terminals left in place.

B= use two identical runs of same wire, termination, length etc. but remove the jumper straps on the speaker and connect high and low with the two different wires.

Now be honest. How many of you are thinking,"Yeah, that would do it." Well guess what? There's a huge design flaw: "B" is using (effectively) doubly thick wire from amp to speaker so the "A" may have an audible disadvantage merely because its wire is half as thick! The only correct way is to use the same double wire for both A and B but fuse it together for "A" and keep the jumpers in place of course.

I wonder if that speaker designer did this? I'll bet "no".
post #79 of 2578
Quote:
Originally Posted by Chu Gai View Post

I also found the following in a USENET posting by Crawford.



Note, that he says there was only ONE test. Now, really!

Hehe, Yes I find it all hard to believe
post #80 of 2578
Quote:
Originally Posted by Chu Gai View Post

So modify it and select the 'No Friggin Idea' button. You still need to run a sufficient amount of trials to achieve statistical signficance. OTOH, you can run what's called a triangle test, where 2 are the same and one is different. All the time. No tricks. Here you make a forced choice as to which is different. If it's strictly due to chance, your success rate is 33%. Done all the time in the food and beverage industry as well as other areas.

NO, you still don't get it! You can't have a 'No Friggin Idea' button or it wouldn't be a DBT by definition! If you really have no 'friggin' idea you'll get roughly 50/50 in the test. Therefore the hypothesis will not be proven. Listeners could actually have more of a 'friggin' idea than they think! They could get 75/25 correct/incorrect by taking a bit of an educated guess each time for example, which might surprise them in the end.

Quote:
Originally Posted by Chu Gai View Post

We rely upon your integrity.

You know I laughed my ass off when I read this. You know if you rely on my integrity, then why won't you let me do a sighted test?
post #81 of 2578
Quote:
Originally Posted by m. zillch View Post

Like you quoted,"The burden of the proof is on the side of the one pretending that a difference can be heard."

Yes that's it, exactly! If there are people here that think Vinyl is better than CD or visa versa, then they need to prove it. But if they don't prove it, the test proved nothing, zilch, zippo, null, zero.
post #82 of 2578
Quote:
Originally Posted by m. zillch View Post

Yes, you are correct, people who lie or cheat during a test make the results worthless. But I'll tell you one way that they can not lie: They can't claim proudly that because they have excellent listening abilities, that they can identify whether X is actually A or B (statistically), when in truth they are unable to demonstrate this ability through A/B/X testing. That kind of lie will be caught by the very test itself. Don't you agree?

Yes I agree. If they can't pass the test, their "lying" hypothesis will not be proven. But if they they can't prove their hypothesis, obviously this does not mean that they were lying. They could have thought they were right about their claims, 100% truthful, but could not prove it in the test. Once again the test proves nothing if failed and that includes that they were lying.
post #83 of 2578
Quote:
Originally Posted by JorgeLopez11 View Post

Classic77.

Don't get confused. DBT are extensively used to make millionaire decisions in food industry, medicine and some other areas where sensory tests are mandatory. They are actually conclusive!

I think Mr Hirvonen's Master Thesis would a be a great help for you to understand ABX tests and subjective audio test methodology.

http://www.acoustics.hut.fi/publicat...rvonen_mst.pdf

Enjoy!

I ain't confused one bit. But you are right about the decision making. DBT's are used to make decisions, like whether bi-wiring in speakers is worthwhile for example. But it does not prove that bi-wiring=single-wiring if the test is failed.
post #84 of 2578
Quote:
Originally Posted by classic77 View Post

DBT's are used to make decisions, like whether bi-wiring in speakers is worthwhile for example. But it does not prove that bi-wiring=single-wiring if the test is failed.

"...because after all, the test subjects may be deliberately sabotaging the results and are selecting A or B at random, even though in truth they can quite clearly hear which is which, but they don't want their tester to know that they can tell the difference."

Is that your point?
post #85 of 2578
Quote:
Originally Posted by m. zillch View Post

"...because after all, the test subjects may be deliberately sabotaging the results and are selecting A or B at random, even though in truth they can quite clearly hear which is which, but they don't want their tester to know that they can tell the difference."

Is that your point?

Well, kinda. For one you don't know whether they are sabotaging the result, so if the test fails no conclusions can be drawn. But lets assume they are not lying, 100% truthful.

Another example, 96kps MP3 vs 97kps MP3. All testers give truthful answers, if they don't we chop off their heads These two codecs are measurably different, we know they are. BUUUUT, maybe because they are so close, the test subjects can only pick X from A and B 51% of the time. The hypothesis that 97kps is better than 96kps has not been proven, we can use mathematics to work out what percentages we need to prove the hypothesis. Does it mean that 97kps=96kps if the test fails? Of course not. Even if we couldn't measure the difference between the two codecs this still would not mean that 97kps=96kps. If the test fails it proves nothing, deosn't matter whether everyone was lying or not.
post #86 of 2578
I'm sure I read somewhere that people can't pick 320kps MP3 from CD in a DBT. In the May 2007 edition of HiFi-Choice magazine, they say the magic number is 224kps (from memory) where humans shouldn't be able to tell the difference. They used some sort of computer program to determine this. None of this is my opinion just what I've read so don't slam me on this guys!
post #87 of 2578
There's a lot of work done on mp3's over at hydrogenaudio.com if'n you want to take a peek.
post #88 of 2578
Quote:


You know if you rely on my integrity, then why won't you let me do a sighted test?

I rely on your integrity but not your biases. Besides, trust your ears!
post #89 of 2578
Quote:
Originally Posted by classic77 View Post

I'm sure I read somewhere that people can't pick 320kps MP3 from CD in a DBT. In the May 2007 edition of HiFi-Choice magazine, they say the magic number is 224kps (from memory) where humans shouldn't be able to tell the difference. They used some sort of computer program to determine this. None of this is my opinion just what I've read so don't slam me on this guys!

Now you are making more sense to me. I'm glad you are in agreement then that we are keeping the liars out of this equation and assume people will honestly do their best to try to distinguish A from B (or whether X is A or B).

The point is if the human ear can't statistically differentiate between A or B then to human perception they sound the same. They may be different in other ways though: They may look different. They may taste different. They may smell different (mmmm polycarbonate ) BUT THEY DON'T SOUND DIFFERENT TO THE HUMAN EAR! They also may or may not measure differently to test instruments. But that doesn't matter.
post #90 of 2578
Quote:
Originally Posted by classic77 View Post

I'm sure I read somewhere that people can't pick 320kps MP3 from CD in a DBT. In the May 2007 edition of HiFi-Choice magazine, they say the magic number is 224kps (from memory) where humans shouldn't be able to tell the difference. They used some sort of computer program to determine this.

Instead of a computer they should have used a properly designed DBT on humans! No, I'm not kidding.
New Posts  All Forums:Forum Nav:
  Return Home
This thread is locked  
AVS › AVS Forum › Audio › CD Players & Dedicated Music Transports › CD vs. Vinyl...the DeathMatch!