Hi-Res 24/94 vs Flac vs CD vs Mp3 files download comparison
Apr 16, 2017 at 8:12 PM Post #138 of 147

christian u

100+ Head-Fier
Joined
Mar 21, 2013
Posts
308
Likes
40
That would be something..... youtube DXD
smily_headphones1.gif
biggrin.gif
rolleyes.gif

 
May 28, 2017 at 12:24 PM Post #142 of 147

sonitus mirus

Headphoneus Supremus
Joined
Mar 23, 2009
Posts
1,849
Likes
653
The problem with ABX is that if I say that I head a difference others will jump on that and say I didn't. So I don't bother.
The Foobar ABX plugin helps with this, but the test is still relying on some integrity from the tester, as there are ways to cheat without too much effort. With the ABX component, the logs use SHA1 hashes for the test files, and the test/log can be verified online. If you provide the Foobar logs with the verification, plus a copy of the test files for others to reproduce the test and analyze the files, it would go a long way to providing some evidence.

Though, I don't know why I even bother, because if you did attempt to ABX two files and the results did not determine that you were hearing a difference, you would surely claim the ABX methodology does not work.

http://www.foobar2000.org/

http://www.foobar2000.org/components/view/foo_abx

http://www.foobar2000.org/abx/signaturecheck
 
May 28, 2017 at 2:32 PM Post #143 of 147

castleofargh

Sound Science Forum Moderator
Joined
Jul 2, 2011
Posts
9,540
Likes
4,940
The problem with ABX is that if I say that I head a difference others will jump on that and say I didn't. So I don't bother.
if you give up on anything as soon as someone has another opinion or asks for more evidence, you won't do much in life. and then sure enough somebody will complain that you're not doing anything so you'd have to stop that too. ^_^
 
May 28, 2017 at 3:34 PM Post #145 of 147

pinnahertz

Headphoneus Supremus
Joined
Mar 11, 2016
Posts
2,072
Likes
738
I won't lie about ABX results. But to have others say I am lying if my results go against your opinion is what causes ABX result not to be posted.
Results of ABX are never "I heard a difference", they are a statistic: number of trials (make this big), number of correct and incorrect matches. Real not all that much to dissagree with. Others can post their results. The only real bone of contention would be the number of trials which affects statistic validity.
 
May 29, 2017 at 10:25 AM Post #146 of 147

Peter Hyatt

Headphoneus Supremus
Joined
Dec 23, 2015
Posts
1,972
Likes
1,914
I don't mind swimming with sharks when I am familiar with the water, but here I have come to learn. I know less than nothing about sound science but respect studies and data (especially when they're explained in non technical language).

With the ABX testing, even at home for self, does it take time (repetition) for the brain to adjust interpretation to the increase of data of lossless over lossy?

Even when I watch an old movie in black and white in low def, after about 30 minutes, it no longer bothers me, as my brain (eyes) adjusts and interprets.

What I heard with a good DAC initially, versus no DAC, was a pleasant discernible difference.

What I heard, 6 months later, after removing said DAC was so much difference that I had no interest in listening.

Is this time to adjust and give brain processing a chance something to consider with A/B testing?

I went to HD Tracks and bought Apple Lossless of "Time Out of Mind."

One track, "Love Sick" is 55.7 mb, 24 bit 1448 kbps

while the CD import of the same (Apple Lossless) is 28.6 mb , 16 bit 744 kbps

Will, over time, the larger file (more data?) mean more brain interpretation?

I appreciate the patience in answering.

Thank you!
 
Last edited by a moderator:
May 29, 2017 at 10:59 PM Post #147 of 147

castleofargh

Sound Science Forum Moderator
Joined
Jul 2, 2011
Posts
9,540
Likes
4,940
like anything else, with experience we can identify some things better. or at least notice something we missed last time. and of course if you can identify some artifact at a particular point in a song, then you can listen to it again and again and learn how to recognize it at that moment and maybe in other tracks. all in all practice can't hurt.
what limits practice is usually just that abx isn't a lot of fun. I believe in it and care for it, even then I get bored real fast ^_^ . you need to set the number of trials before starting(it's important for statistical reasons), and I have to admit that getting to the end can be a struggle for me sometimes ad I will stop, then do the second half the next day. which is very fine if you feel like you're getting some sort of fatigue. nothing force you to do all your trials in one go you do 2 per week if it pleases you. but try to stick to the predetermined number of trials.

what you say about the human brain and its ability to adapt to pretty much anything is true. we constantly discard so much of the audio signal we receive, because it's interpreted as less important, because we weren't focused on that particular instrument at the time, because it's some reverb or room acoustic and the brain is super good at kind of removing it or interpreting it as being a wall or a room instead of the music. also memory plays some tricks. in fact it is the very reason why we need stuff like ABX, to reduce the delay between 2 samples so that memory or brain interpretation don't have time to alter the event too much. rapid switching works great that way and increases the chance that if we feel something different, it is. then we repeat to make it a statistical result and push random luck out of the way.
as for how you will proceed when doing abx tests, you can decide to test different conditions yourself. wait longer before switching back, listen to long samples or use a very short passage in a loop until you're sick of hearing it. in some cases what you're trying to identify will condition the test method. but in general, short samples should work nicely because of how the memory works. the difficulty then might be to judiciously select that short sample within an entire song.

about the files, again it depends on what question you're trying to answer, but if you're testing the formats and only the formats, then you will want to take the high res version and convert it yourself(people here can provide advice if you need them). that way at least you know for sure that the files all come from the same origin. usually we suggest to convert down to the lower format, and then convert that back to the high res resolution. this is done so that what you listen to is only the data variation and not the way your DAC or your computer deals with different formats.

all in all checking and double checking is always a good idea, after all we're involved in a test to prove repeatability. so if you get lucky and find tracks with audible artifacts, you might share them(samples only for copyright reasons), so that other members interested can also try to replicate your results and participate in gaining confidence in the results. or maybe nobody else will get your result and it might suggest a problem in your testing method, you gear, or that you're a golden ear ^_^.


if I said something stupid, a fellow member will point it out. I'm after all far from being an ABX expert. I've done a lot of them but in a personal and fairly lenient way when it didn't feel very important to me.
 

Users who are viewing this thread

Top