Skip to main content
Topic: Does ABXing account for all factors? (Read 2771 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Does ABXing account for all factors?

When people speak about compressed audio being fatiguing to listen to, is this something that can be ABXed? 

It seems that a genuinly audible difference produced by compression could be ABXed.  But if ABXing is usually done with relatively short clips of music a person's ears might not have time to become fatigued. 

Is ABXing missing elements that are not initially noticable when listing to an audio clip, but that become more apparent after prolonged listening?

Ian

Does ABXing account for all factors?

Reply #1
<gullible>OMG!! No wonder I've been so tired lately! It must have been those fatiguing lossy files I've been listening to so often!</gullible>

I suppose you might get tired of "listening for artifacts," so that would mean you'd get less likely to notice artifacts after prolonged listening.

That's why ABX focusses on short samples; to lessen listening fatigue.

Does ABXing account for all factors?

Reply #2
Quote
When people speak about compressed audio being fatiguing to listen to, is this something that can be ABXed?

When people speak of being fatigued, they are usually referening to the effort it takes, and the strain placed on your ears, when do an ABX test.  It's not an element to ABX, just a byproduct of performing a test, the person doing so becomes fatigued.

Quote
Is ABXing missing elements that are not initially noticable when listing to an audio clip, but that become more apparent after prolonged listening?

In my opionion that is unlikely, but I don't know of any concrete evidence one way of the other.

Does ABXing account for all factors?

Reply #3
When I bought a cheap Phillips carousel CD player, my original plan was just to use its transport and keep all playback through my fancy external DAC. I didn't hear any difference at all comparing the outputs, so I got lazy and just used the Phillips analog outputs. Over about a month I noticed I wasn't listening to CDs as much as I used to, and that made me very concerned that some kind of longer exposure effect might be causing me fatigue without raising a pointy head and being obvious about it.

The essential part of ABX is that it is double blind, how long you take to decide if x is A or B is kind of up to you, take 10 seconds or 3 months, just keep it double blind and do a reasonable number of trials.

BTW I figure audiophiles are meant to suffer, so instead of changing to my old DAC, I am shopping for a new amp.

Does ABXing account for all factors?

Reply #4
There's been some talk before about doing long-term ABX tests. A couple of real-world examples:

The test that was performed by some scientist  that showed people could tell when ultrasonic frequencies were played along with the 20Hz-20kHz audio. I think the ultrasonic tweeters were turned on for durations of at least one minute.

The 24-bit audio test that was prepared by some forum. ff123 talked about it, so if you want to find out the test, search for posts by him. Anyway, the guy who passed the test did one ABX test per day over the course of a week.

Pio2001 has also mentioned that, for one of the ABX-able MPC samples he found, he did much better when he let his ears rest for a minute or more between playbacks.

There's ample evidence that quick successions of ABXing can reduce your sensitivity to artifacts. There have also been suggestions (perhaps by 2Bdecided, among other forum members) that users try long-term tests to see if lossy audio is more fatiguing. Basically, you keep two copies of each song in your playlist: one lossless, one lossy. Make sure they look identical in the playlist (same formatting string) and have the same volume (use Replaygain). You'll have to randomize their order somehow, too. Then, whenever you're chilling out and listening to your tunes, and you suddenly think to yourself "this must be the lossy version" or "this has gotta be the lossless version", look up the filename of the currently-playing song and see if you were right. Tally up your correct and incorrect guesses, and after you've accumulated a pile of guesses (over the course of days or even weeks), check the p-val of your long-term ABX test.

Quote
The essential part of ABX is that it is double blind, how long you take to decide if x is A or B is kind of up to you, take 10 seconds or 3 months, just keep it double blind and do a reasonable number of trials.

Right on, MikeFord! And welcome to Hydrogenaudio!

 
SimplePortal 1.0.0 RC1 © 2008-2019