Head-Fi.org › Forums › Equipment Forums › Sound Science › An experiment in lossy vs lossless A/B testing
New Posts  All Forums:Forum Nav:

An experiment in lossy vs lossless A/B testing - Page 2

post #16 of 51
Warning: Spoiler! (Click to show)
Be been watching this discussion and I'm glad it has its own thread now. This hits pretty close to home for me because I'm currently in the process of rebuilding my music library and I'm wondering if its worth going to flac or sticking with MP3. I'm on my tablet so I just set the two to play one after the other on loop. They both sound pretty close to me but putting my cared on the line I'm going to say b is the flac. It sounds a bit fuller and more nuanced in the guitar and has a more natural decay to the chimes.
Could everyone who partakes please list their test equipment? I'm still pretty budget fi so I know my equipment isn't as revealing as some. I'm hoping higher end equipment doesn't make it easier to tell the difference,
I am using a PiPO s2 android tablet with jetAudio+ and my monoprice 8320 and Sony mdr-v7
Edited by thoughtcriminal - 2/8/13 at 4:46pm
post #17 of 51

Completely, utterly disregarding whether anyone can or can't hear a difference, why not use lossless only for the fact that there is no loss? Then use it to convert down to whatever MP3 you feel comfortable with, be it 64, 128, 256, or whatever. Storage is very cheap these days. A $130 3TB hard drive can store, probably, all the music you'll ever own. 

 

My current testing equipment is HiFiMan HE-400 and Etymotic ER6i (technically a rebrand) played through a PCM2704->CMOY combo.

 

Also, please hide with the "spoiler" tag any of your impressions or conclusions. Cheers!

 

Quote:
Originally Posted by thoughtcriminal View Post

Be been watching this discussion and I'm glad it has its own thread now. This hits pretty close to home for me because I'm currently in the process of rebuilding my music library and I'm wondering if its worth going to flac or sticking with MP3. I'm on my tablet so I just set the two to play one after the other on loop. They both sound pretty close to me but ...........
Could everyone who partakes please list their test equipment? I'm still pretty budget fi so I know my equipment isn't as revealing as some. I'm hoping higher end equipment doesn't make it easier to tell the difference,
I am using a PiPO s2 android tablet with jetAudio+ and my monoprice 8320 and Sony mdr-v7

Edited by soundstige - 2/8/13 at 4:45pm
post #18 of 51
Thread Starter 
Quote:
Originally Posted by soundstige View Post

Completely, utterly disregarding whether anyone can or can't hear a difference, why not use lossless only for the fact that there is no loss?

 

I'm confused. You're obviously 100% set on the idea that the difference is imperceivable so why waste the disk space? Sure, "there is no loss", but according to you that makes absolutely no difference at all. So why are you advocating lossless files? For no reason whatsoever?

 

BTW, if we're doing 2 wav files where we don't know which one is the lossy copy then the foobar ABX component isn't neccessary and your ABX log tells us nothing at all because it doesn't tell us which file you chose.


Edited by devhen - 2/8/13 at 4:58pm
post #19 of 51
Quote:
Originally Posted by soundstige View Post

Completely, utterly disregarding whether anyone can or can't hear a difference, why not use lossless only for the fact that there is no loss? Then use it to convert down to whatever MP3 you feel comfortable with, be it 64, 128, 256, or whatever. Storage is very cheap these days. A $130 3TB hard drive can store, probably, all the music you'll ever own. 

 

My current testing equipment is HiFiMan HE-400 and Etymotic ER6i (technically a rebrand) played through a PCM2704->CMOY combo.

 

Also, please hide with the "spoiler" tag any of your impressions or conclusions. Cheers!

 

Hidden! Thanks! Ive lost a lot of music to crashing hard drives, easily more than 10000 songs. Now im paranoid and will only use multiple layers of redundancy. henceforth all my music will reside on my microsd cards as well as two hard drives and possibly dvd backup. This adds up quickly when youre a young man just trying to figure out how to make it in this world


Edited by thoughtcriminal - 2/8/13 at 5:05pm
post #20 of 51
Quote:
Originally Posted by devhen View Post

 

I'm confused. You're obviously 100% set on the idea that the difference is imperceivable so why waste the disk space? Sure, "there is no loss", but according to you that makes absolutely no difference at all. So why are you advocating lossless files? For no reason whatsoever?

 

BTW, if we're doing 2 wav files where we don't know which one is the lossy copy then the foobar ABX component isn't neccessary and your ABX log tells us nothing at all because it doesn't tell us which file you chose.

You are misstating his argument for like the tenth time. :D

post #21 of 51

You do know all you're doing is showing you have an axe to grind against me by assuming what I think, right? A bias. My whole 200+ CD collection is in FLAC. Come on, don't be so one-dimensional, we're all just trying to enjoy a test here. Especially don't be so one-dimensional when that one dimension was created solely from a misunderstanding of my points...

 

Quote:
Originally Posted by devhen View Post

 

I'm confused. You're obviously 100% set on the idea that the difference is imperceivable so why waste the disk space? Sure, "there is no loss", but according to you that makes absolutely no difference at all. So why are you advocating lossless files? For no reason whatsoever?

post #22 of 51
Thread Starter 

So whats the argument? Because first it was that he wanted "proof". Then I did an ABX and he instantly called me a liar because the results I got didn't match his predetermined position. So maybe you can explain to me his argument because I read everything he wrote in the other thread and all I heard was a guy trying every angle he could possibly come up with to claim that the difference is impercievable and people who think otherwise are mistaken.

post #23 of 51
Quote:
Originally Posted by devhen View Post

 

I'm confused. You're obviously 100% set on the idea that the difference is imperceivable so why waste the disk space? Sure, "there is no loss", but according to you that makes absolutely no difference at all. So why are you advocating lossless files? For no reason whatsoever?

 

BTW, if we're doing 2 wav files where we don't know which one is the lossy copy then the foobar ABX component isn't neccessary and your ABX log tells us nothing at all because it doesn't tell us which file you chose.

There's still good reason to have lossless for archiving purposes if you can't tell the difference. Then there's always peace of mind.

 

The ABX log says you can tell a difference. Just posting your answer shows that you could have been guessing, it's only a 50% chance after all. 

 

These tests usually don't blow up enough to really hold any statistical significance, but I think it's good for people to see for themselves what the different types truly are back to back.

post #24 of 51
Thread Starter 

Look I'm not trying to start another argument. I'm genuinely curious. You can't say that your position is that its extremely rare to be able to tell the difference so it must be proven but then when people try to "prove" it to you (i.e. explain their experience to you) you automatically shoot everything they say down. It doesn't work that way. Either you want to analyze all possible evidence in a fair way or you want to specifically discredit everyone and everything that supports the theory that the difference is noticeable.

post #25 of 51
Thread Starter 

That last post was for soundstige.

post #26 of 51
Thread Starter 
Quote:
Originally Posted by chewy4 View Post

There's still good reason to have lossless for archiving purposes if you can't tell the difference. Then there's always peace of mind.

 

The ABX log says you can tell a difference. Just posting your answer shows that you could have been guessing, it's only a 50% chance after all. 

 

These tests usually don't blow up enough to really hold any statistical significance, but I think it's good for people to see for themselves what the different types truly are back to back.

 

Am I misunderstanding this? If we don't know which file is which then neither does the ABX component and the only thing it tells us is if we guessed right. How does that help if neither us or the ABX component knows which file is which? soundstige's ABX log doesn't reveal anything about which file he chose as the lossless version.

post #27 of 51
Quote:
Originally Posted by devhen View Post

 

Am I misunderstanding this? If we don't know which file is which then neither does the ABX component and the only thing it tells us is if we guessed right. How does that help if neither us or the ABX component knows which file is which? soundstige's ABX log doesn't reveal anything about which file he chose as the lossless version.

Well it's still good to say if you think A or B is lossless, but ABX lets you know if it's a guess or not.

post #28 of 51

I will take you 100% seriously:

===============

Quote:
"You can't say that your position is that its extremely rare to be able to tell the difference so it must be proven but then when people try to "prove" it to you (i.e. explain their experience to you) you automatically shoot everything they say down."

I never did this.

 

Quote:
Either you want to analyze all possible evidence in a fair way

Yes, that's what I want to do.

 

Quote:
or you want to specifically discredit everyone and everything that supports the theory that the difference is noticeable.

That sounds like a silly waste of time, why would I do that? If you at all think that's what I've been trying to do, ever, I'm sure that's the reason you're not yet in agreement with me. All I've stated is that being able to tell the difference is very difficult and rare, and requires proof.

 

=================

Quote:
Originally Posted by devhen View Post

Look I'm not trying to start another argument. I'm genuinely curious. You can't say that your position is that its extremely rare to be able to tell the difference so it must be proven but then when people try to "prove" it to you (i.e. explain their experience to you) you automatically shoot everything they say down. It doesn't work that way. Either you want to analyze all possible evidence in a fair way or you want to specifically discredit everyone and everything that supports the theory that the difference is noticeable.

post #29 of 51
Thread Starter 

Yes and you instantly accused me of lying when I posted my initial results. Can't you see that that that would make me feel like you're biased?

post #30 of 51

I didn't say anything to your initial results. You never provided results or evidence of any kind. You said that you got 5/5 right. That's not proof, that's hearsay. 

 

When you post a proper ABX log, like I've already done in this thread, preferably of at least 10 trials (I should have done more than 7), then we can make conclusions from that.


That's the initial time you'll have provided evidence. I'm sorry if 5/5 were your actual results and I didn't believe you, but a lack of actual evidence combined with constant insults from you gave me no reason to believe.

 

Your opinion/hearsay doesn't matter. My opinion/hearsay doesn't matter. We are equals. Statistics matter. I have provided statistics (albeit not as good as I could have). You have so far provided zero statistical evidence. I would like you to do so.

Quote:
Originally Posted by devhen View Post

Yes and you instantly accused me of lying when I posted my initial results. Can't you see that that that would make me feel like you're biased?

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › An experiment in lossy vs lossless A/B testing