http://www.head-fi.org/forums/showthread.php?t=269401

Okay that was just for entertaining reading and it could put this into context, but...

The real point of this thread is to ask whether there is any possible technical difference between Foobar's bit-comparator and ABX comparator. In other words, if the bit-comparator says there are no differences between the output WAV, is the ABX comparator also using identical WAVs? Or does it use some different pipeline, opening up the slight possibility that there could be a difference between the ABX files, owing to some bug, but not between the bit-comparator files?

Question relevent to all versions but do comment concerning any differences.

Moderation: Adjusted the topic title, changed the link to point to the components site - direct links won't work.

I'm not gonna believe something like this until i see him doing ABX myself. especially someone who saying that 32bit wav are sound better than 16bit come from 16bit source. AND especially from head-fi.

That said. His testing is flawed. he use different source to abx. I don't see him abx WAV & lossless from the same wave file. but then again. I bet he will tell that "wave sound better" anyway and there is no log. no statistic, nothing but what he seem to feel in what he said. is this really "ABX" ?

Doesn't matter what the source was eh?

AIFF & ALAC -> ABX means the ABX comparator writes out two WAVs for the comparison.
AIFF -> WAV
ALAC -> WAV

So does the bit-comparator, and that says the resulting WAVs are identical (at least for me it does - the guy in question refuses to do the test). So, my conclusion is that there could not possibly be a difference in the ABX either.

I don't believe the guy either, but just want to interrogate my own conclusion.

Actually it just occurred to me, where does Foobar stick the temporary WAVs used for the ABX? Seems that they could just be examined to answer this question. Although it would still leave open the question to versions other than mine.

I don't think there's any point insulting him. I walk past hundreds of people of dubious intelligence every time I walk through town - I don't feel the need to tell them how stupid they are (and if I do, I tend to restrain myself for politeness and self -preservation!).


Of course it's ridiculous. He claims to hear a difference when playing two identical wav files. When this is pointed out, he says he understands that they're identical - but they might be different because they were converted from different sources. Bit identical but different! Well, certainly stored on different parts of the HDD, but if that wrecks audio quality, we are all stuffed!

If you push him on this, when he finally understands that there really can be no consistent difference, he'll retract the ABX results, saying that it was just luck, and (he's going this way already) harder to hear and more random under ABX then when playing the actual FLAC/WAV/whatever, which really do exhibit an audible difference.

You'll be left in the usual situation: there could be something wrong, but only in the circumstances where it's entirely impossible to prove otherwise, for reasons that make no sense, and in ways that can never be tested (or the OP refuses to test).

The sad thing is, occasionally there are bugs found in this way: the MAD mp3 decoder used to add clipping, but only to existing clipping - someone found this just by listening and wasn't believed. Thankfully the objective evidence was forthcoming. A combination of ignorance, mindset, and (I hate to say this your aggressive insults mean any possible evidence isn't likely to come from this discussion.

Cheers,
David.

As both tests use the same decoder, there should be no differences between the WAVs bit-compared and the WAVs used in the ABX test.
However, you can intentionally make them differ. Take two identical files and ReplayGain-scan one of them. The WAVs in the bit-comparison will still be the same, but the ABX test WAVs can be different if the use of ReplayGain is enabled.

Thanks David... though if you just read the thread recently, a Head Fi mod has gone through it and deleted my posts, removing the more constructive ones and apparently missing one or two which happen to be more colorful. I don't have much time for idiots like this, as it just seeds the same sort of nonsense in newbies and I see no reason to adhere to political correctness. If he's gonna post it then why not have him on. He's not innocently naive...

Does anyone know the answer to my core question, namely do the ABX and bit comparators compare the "same" WAVs? Is there any possible technical reason that the two ABX WAVs could differ? This answer will it seems conclude the larger issue.


You mean if the Use ReplayGain option is checked in the ABX start dialog?

Okay I think that answers the question.
Got any idea where Foobar sticks the temp WAVs for ABX so as to verify?

As a matter of fact, I just did verify it. I started an ABX test between two identical files, then browsed to the temp directory (foobar tells you where that is at the beginning of an ABX test), added the temporary ABX WAVs to the playlist and bit-compared them, first to each other, then to the source files.

No differences.

Rock n roll

Well that clears it up, only remaining possibility is that his version of Foobar is somehow different, but since he refuses to do that same examination (how long does it take eh?) I guess we'll just have to each make our own assumption...