Gather around folks for my new training for "How to hear the difference and ace an ABX test". We all know how important training is right? But unlike my last two methods of passing the ABX test, by listening alone and not cheating, where I refused
to share the picky details due to my fear it would fall into the wrong hands and be used for fraud, NOW I'M GOING TO SPILL THE BEANS
since I've already documented that the current versions of Mosaic A2 and B2 have a 10 to 11 millisecond delay between them, easily seen on my previously posted screen capture images.
Acoustical memory is short and easily distracted by exposure to unrelated material. Instead of focusing on the difference at the beginning of this short guitar section, focus on the sound of the last strings getting plucked at the end. These are the exact same two time sequences taken from Mosaic B2 and A2. I could have made the segments longer, say 10-20 seconds or even more, and the difference in the guitar pluck sound at the end would have been identical, but your memory
for how it sounds wouldn't have been as intact.
Listen to the two for awhile and eventually you can hear a difference without even needing to alternate between the "Play A" and "Play B" buttons in foobar ABX [or whatever sighted comparison program you like using]; all you have to do is hit "Play X" in ABX and you can tell which one it is straight away without any assistance! Woo Hoo!
I just passed this test in record time, two minutes, no switching midstream between A and B at all,
and in fact in many of my trials I never even had to hit any buttons except "Play X", I could tell which it was without even needing a quick refresher as to the sound of A or B!
Here's my very first attempt at ABXing them, planning ahead of time to do only 11 trials total, no matter how well I was doing, to make krabapple happy
[He's right though.]:
foo_abx 1.3.4 report
File A: C:\Users\Me\Desktop\Mosaic_A2 excerpt taken from 1m52s.wav
File B: C:\Users\Me\Desktop\Mosaic_B2 excerpt taken from 1m52s.wav
12:40:10 : Test started.
12:40:27 : 01/01 50.0%
12:40:43 : 02/02 25.0%
12:40:52 : 03/03 12.5%
12:41:12 : 04/04 6.3%
12:41:29 : 05/05 3.1%
12:41:39 : 06/06 1.6%
12:41:49 : 07/07 0.8%
12:41:58 : 08/08 0.4%
12:42:09 : 09/09 0.2%
12:42:25 : 10/10 0.1%
12:42:34 : 11/11 0.0%
12:42:36 : Test finished.
Total: 11/11 (0.0%)
I recommend listening to this short guitar excerpt from Mosaic B2 first, get acquainted with it, and then
listen to the sonic character of the final guitar pluck sound at the very end of Mosaic A2:
[Go to this folder, then right
click on each of the two .wav files it contains, to then select the download option]:
Do you hear it? I picked the same section I showed in my earlier screen shot so it would be easier to hear, but in theory such a difference may occur all over the place. I just took a stab at where I thought it would be prominent, and I was right, first try. Many people have mentioned they were focusing on the dynamics, too, so selecting such peaks is to be expected.
Don't fall for people saying, "Oh no, but it's not that
I'm sensing at all." All they know is they are hearing some
difference and it's human nature to attribute it to all sorts of other things than what it actually is. As JJ pointed out we know this from small < 1dB level changes; people almost always attribute it to "clarity", "detail", "three dimensionality", etc. instead of what it really is. It is just the way humans are; no reason to be ashamed of it, it's true of everyone.
Here's the take away moral to the story:
ANYONE WHO LISTENS TO A SEGMENT OF THESE SONGS, RATHER THAN THE ENTIRE SONG START TO END WITH NO SWITCHING MIDSTREAM, MAY VERY WELL BE SENSING EXACTLY WHAT I HAVE ILLUSTRATED TO YOU HERE, BUT THEY MAY NOT CONSCIOUSLY REALIZE IT IS WHAT THEY ARE KEYING OFF OF!
The reason the guitar pluck sounds different is because 10 milliseconds have been sliced off due to these files' time alignment problem and everyone using the segment selection option, which if I recall correctly everyone said they did
, would similarly be having something sliced off as well.
Also don't fall for anyone telling you 10 milliseconds is too small to make any audible difference: YOU JUST HEARD IT YOURSELF!
[Note: I'm new to Audacity and I selected this short passage by eye and by hand. There may be more precise methods but I haven't learned them yet.]