Do you ABX? What can you ABX? Do you care?

[NotJeffBuckley]

Run a CRC verification and you'll have your answer as to whether the copy is perfect or not. Bits are bits, and there's no point adopting some sort of strange holistic viewpoint of the process. Compound degradations also don't come into play unless you're copying copies for many, many, many generations, and if each successive copy is successful (has no CRC errors when recompiled) then the successive copies from that one will be, too.

CRC every copy for 24, or any other arbitrary number of copies, and if each matches the original you've had no degradation.

Edit: And, yes, I ABX and DBT. It's the only way to remove the observation variables and get a true analysis. It's letting your ears tell you the truth without interference from your eyes and the expectations that come with expense.

 

[meat01]

Quote:

Music to me is for relaxation and pleasure, not hard work.

I agree, but it seems everyone strives for the nuances and pays hundreds of dollars for that last percent of audio quality. If something is not easily distinguishable blind, then it isn't worth the money in my opinion.

 

[kwtoxman]

Someone please give a brief synopsis of ABX.

Here is my story. Being an analyst/scientist by profession I have a hard time believing that 99.99% of people could tell the difference from an original to copy cd (no mp3 compression). And I think true testing (double blind, I assume ABX, etc) is the only way to find out.

I have an audiophile friend with a very nice system who did subscribe to the theory that he could tell the difference between copied and original CD's. So on one afternoon I set up a session where I played an original cd of his (he knew it very well) and a copy on his system. The cd changing was totally randomized so that at times he would blindly listen to the original 2 times in a row, rather than always going between original and copy. After acclimating himself to 6 listening episodes (20 second episodes of a familar track) he had to decide the next 4 episodes. Were they original or copy. He got the first assessment correct, and the next three wrong. There was a single back to back same disc, and he assessed a difference.

While not technically enough samples for a statistically significant outcome, it was very enlightening to him.

kw.............

 

[Senn20]

I value my sanity, so I choose not to ABX test.

 

[bordins]

From Computer-Driven Audio - Is it Superior to Optical-Based CD Playback?

"... This was demonstrated to me by Mark Hampton of Zcable, who created some excellent CD-R's and shared some of them at 2005 CES in Las Vegas. The improvement was not subtle. Several other manufacturers are duplicating CDs on CD-Rs for their own use, and they even insist on using special CD-R media. ..."

If you don't believe the above quotes, you need to test it yourself. Proper ABX tests allow us to remove all personal biases (subjective opinions/beliefs) out of the test.

 

[fewtch]

Quote:

Originally Posted by bordins If you don't believe the above quotes, you need to test it yourself. Proper ABX tests allow us to remove all personal biases (subjective opinions/beliefs) out of the test.

Is there something wrong with subjective opinions/beliefs? I guess objectivists seem to think so, as they whine constantly about how other people choose to spend their time/money.

 

[bordins]

Quote:

Originally Posted by fewtch Is there something wrong with subjective opinions/beliefs? I guess objectivists seem to think so, as they whine constantly about how other people choose to spend their time/money.

The fact is ABX is very solid and there is nothing wrong with personal preferences. Choose whatever method you like, but ABX is still "proven".

 

[JJ15k]

I think that after having spent xxxx dollars on highend audio gear, a lot of people wouldn t want to realise that they can t see the difference between a 1000$ and a 150$ piece of equipment.
Personnally I d love to see more ABX tests so I could choose my gear accordingly, however we re on a forum so I beleive that s too much to ask(I m already grateful to be able to hear subjective opinions).
But a real audio reviewer should do one...except they don t since otherwise they couldn t push expensive products

 

[Chri5peed]

Quote:

Originally Posted by NotJeffBuckley Run a CRC verification and you'll have your answer as to whether the copy is perfect or not. Bits are bits, and there's no point adopting some sort of strange holistic viewpoint of the process. Compound degradations also don't come into play unless you're copying copies for many, many, many generations, and if each successive copy is successful (has no CRC errors when recompiled) then the successive copies from that one will be, too. CRC every copy for 24, or any other arbitrary number of copies, and if each matches the original you've had no degradation.

Completely true, also add in the fact that errors copying files on harddrives are as rare as gold-dust these days.
Anyway, if I make a copy of audio files I keep the original files and use them to copy from anytime.

 

[darkless]

As long as the original audio is stored as regular files and not as audio tracks on a CD, you can make an infinite number of copies with no "generation loss", since regular file copying is 100% error proof. Period.

 

[Hirsch]

Quote:

Originally Posted by bordins If you don't believe the above quotes, you need to test it yourself. Proper ABX tests allow us to remove all personal biases (subjective opinions/beliefs) out of the test.

Does it? You're subscribing to the theory that part of the perceived percetual difference is an expectancy effect, that is causing perceived differences where none is present. If expectancy effects can cause "false alarms", or perceived differences where none are really present, how do you deal with "misses"? If you go into an ABX test believing it will show no difference, what do those same expectancy effects predict? (Hint: expectancy effects predict that scepticism about differences should make real differences less detectable.) Does the ABX test include positive controls, where there are known and measurable differences that should be perceptable as a fundamental part of the process? (Hint: It does not.) In the absence of those, the sensitivity of the test is unproven, and a negative result becomes uninterpretable.

 

[NotJeffBuckley]

Quote:

Originally Posted by fewtch Is there something wrong with subjective opinions/beliefs? I guess objectivists seem to think so, as they whine constantly about how other people choose to spend their time/money.

I don't see anyone whining. Kickin' strawman!

And yes, for the purposes of absolute variance evaluation, there is something wrong with subjective opinion/belief. However, for personal listening pleasure, nothing wrong with it at all.

 

[NotJeffBuckley]

Quote:

Originally Posted by Hirsch Does it? You're subscribing to the theory that part of the perceived percetual difference is an expectancy effect, that is causing perceived differences where none is present. If expectancy effects can cause "false alarms", or perceived differences where none are really present, how do you deal with "misses"? If you go into an ABX test believing it will show no difference, what do those same expectancy effects predict? (Hint: expectancy effects predict that scepticism about differences should make real differences less detectable.) Does the ABX test include positive controls, where there are known and measurable differences that should be perceptable as a fundamental part of the process? (Hint: It does not.) In the absence of those, the sensitivity of the test is unproven, and a negative result becomes uninterpretable.

ABX can show indeterminability just as well as determinability. Further, "skepticism about differences" assumes a positive desire to see no result, which anyone performing an ABX shouldn't have; there's no point trying to prove the null hypothesis, and ABXing takes time and effort to do.

 

[Publius]

Quote:

Originally Posted by NotJeffBuckley ABX can show indeterminability just as well as determinability.

Actually, no it doesn't. You need to make a subjective judgement call to interpret a negative test result as indicating no real difference. Granted, given low enough test error, and given that one agrees with the quality of the testers, a lot of people will do that.

 

[bordins]

It would be relevant to mention about studies performed in engineering disciplines to determine how well machine controls, for human operators, are. For example, Patriot missile radar screen, airplan cockpit, car dashboard, etc.

The machines are designed to aid human judgement in determining signals of interest among the presence of noise (including no signal). The test is the signal detection task. There are a lot of theories under signal detection theory. How good a new design is *DEPENDS* on *PERFORMANCE* of human perception (ability to identify). Whether or not, the operator expects to have any difference.

Given a good experimental design, the *RESULTS* can show you whether there exists such improvement. This is merely reflected by human performance.

In ABX, the task is to identify any sonic difference through the loudspeakers. You may find only exceptional people, with very-well-trained ears, can correctly identify differences between two designs (audio components). But how about ordinary people even youself. ?

This is also similar to how good a pain-killer pill is. Take a placebo v.s. the pill without knowing which you take. Make you some pain and let you rate the pain level. The test is subjective in nature, but the result is objective. Try some pills you have at home.