Head-Fi.org › Forums › Equipment Forums › Sound Science › Double blind test 128Kbps vs lossless? I'll be amazed if you can tell much difference
New Posts  All Forums:Forum Nav:

Double blind test 128Kbps vs lossless? I'll be amazed if you can tell much difference - Page 4

post #46 of 257
Quote:
Originally Posted by Br777 View Post

touche.

 

I guess I am clearly not up to date on how far audio encoders have come.    the last time i heard a 128 file was probably early 2000 and i can still remember how awefull it sounded. 

having only that reference for "low quality" files, its still hard for me to believe that the compressed file is 128... but hey.. it probably is, and if thats the case... more power to it.  Its amazing how far file compression technology has come.


x2  My experience 10 years ago w/ MP3 caused me to go WAV and never look back.  Took a lot to convince me to go lossless even after research and checksum analysis.  I still have some nagging questions and experiences about data transfer and encoding for sound quality but I admit my paranoia probably verges on the edge of voodoo and black magic.

post #47 of 257

1 was probably the 128 MP3 file. Technics CD player -> M^3 amb amp -> HD 800

post #48 of 257
Quote:
Originally Posted by beamthegreat View Post

1 was probably the 128 MP3 file. Technics CD player -> M^3 amb amp -> HD 800


Just to throw it into the mix before anyone would assume so, I would like to add that it doesn't take an astounding setup to tell the difference (and not to take away from those who do have a much, much better setup than me either. I'm a poor college student, it's going to be a little lacking). But my setup for this test was Lenovo R61i thinkpad headpone out > FiiO E5 > Head Direct RE2.

post #49 of 257

Huge difference. You really gotta pick it up right away, or else you will listen to carefully (if that makes sense) and you will trick yourself into hearing the better one. Number 2 definitely has more body to it. Setup: ZeroDAC>MisterX M3>Denon D5000

 

If I am incorrect, I will be shocked. 

post #50 of 257
Quote:
Originally Posted by stang View Post

Huge difference. You really gotta pick it up right away, or else you will listen to carefully (if that makes sense) and you will trick yourself into hearing the better one. Number 2 definitely has more body to it. Setup: ZeroDAC>MisterX M3>Denon D5000

 

If I am incorrect, I will be shocked. 


what have you recently shopped.
 

post #51 of 257

Traded HF-2 for D5000 for a couple of weeks  They are about as bright as the Grado's from memory. Which annoys me, but I am now used to D5000 treble and I enjoy them. Lots and lots of bass though...don't really like it. 

 

Lol back on topic now...

post #52 of 257

I could find difference between 225kbps and flac on my speaker rig and it doesn't have to be even playing the same track in different formats.play the track and I will straighaway find out what format it is.This was listening to house and trance go figure what's possible with my headphone rig.

post #53 of 257
Quote:
Originally Posted by stang View Post

Traded HF-2 for D5000 for a couple of weeks  They are about as bright as the Grado's from memory. Which annoys me, but I am now used to D5000 treble and I enjoy them. Lots and lots of bass though...don't really like it. 

 

Lol back on topic now...



I would like to give you some advice, please don't get the DAC1 way over priced for the content, get Lavry.

post #54 of 257
Quote:
Originally Posted by kunalraiker View Post





I would like to give you some advice, please don't get the DAC1 way over priced for the content, get Lavry.


I know. Not buying it. Possibly trading some of my gear for it: HF-2 and maybe something else. Getting it from a friend who doesn't use it + some other stuff of his possibly. If it wasn't for this, I would try buy something like a DIY Buffalo DAC (already made ofc). He really wants the HF-2 and I don't use them anyway. Too bright. Maybe I can't stand low ohm headphones anymore without getting used to them and EQing them? 

 

Lol thread derail 

post #55 of 257
Quote:
Originally Posted by stang View Post

Huge difference. You really gotta pick it up right away, or else you will listen to carefully (if that makes sense) and you will trick yourself into hearing the better one. Number 2 definitely has more body to it. Setup: ZeroDAC>MisterX M3>Denon D5000

 

If I am incorrect, I will be shocked. 

You may have a point there. When I first listened to the tracks, my immediate inclination was to say that track 1 was the lossy version. But then I wanted to listen closely to be sure, and I became far less sure than I first felt. I still came out with the correct result, but it ended up taking me far longer than it would have if I had trusted my first inclination. (And I guess with so many people giving there results now, no real point in trying to avoid creating a bias for others any more)
 

post #56 of 257

Uhh, you guys know how to to a blind test right? You need the foobar ABX comparator plugin, else it's not really blind imo. You need to use replaygain to level match it, there's usually a slight volume difference between lossless and mp3.

Quote:
Originally Posted by Br777 View Post

 

so....music sounding good to me both compressed and lossless makes me a victim of snake oil??? 

 

where is the context for that comment?  I dont get it.

 

 

 


 


My point is that headfi members shill things and nobody takes blind tests or measurements to confirm if it's actually better. One thing that strikes me as blatantly obvious is that people on headfi are rich and plug things based on price and not an objective or measurable verification of the hardware's accuracy.

 

This is especially true for DACs and headphone amps. Level matched, I think that most people here won't tell a difference in a blind test. Plus, people try to sell amps where it's not needed.

 

Some companies like Audio GD likes to sell "colored" gear for example, if you mail them asking for measurements they will ignore you. The only RMAA measurements I've seen of their stuff was certainly not flattering. Same with distortion measurements of their opamps.

 

I had asked the same thing to HA-info, and the results were much the same. Won't give info. Advertised ">120 db snr" had translated into a real life performance of ~100db snr with a bad frequency response... and to top it off, the Chinese Build Quality(tm) had left me with a piece of junk that also would cut out the left channel, or work at low volume on both channels until you fidget with it for half an hour.


Edited by Dalamar - 7/6/10 at 12:36pm
post #57 of 257


 

Quote:
Originally Posted by Dalamar View Post

Uhh, you guys know how to to a blind test right? You need the foobar ABX comparator plugin, else it's not really blind imo. You need to use replaygain to level match it, there's usually a slight volume difference between lossless and mp3.

Actually, using the ABX plugin would only be important if you knew which track was which going into this. Since we're already prevented from knowing which track is lossy and which is lossless just by looking at the track, using the ABX plugin in this scenario would only add in confounding variables since you then not only have to figure out whether track 1 or track 2 is lossless, but also whether track X or track Y is the same as track A or track B which creates much more opportunity for you to trick yourself into hearing the same thing, audio memory being what it is. It would be an incredibly poor way to set up an experiment.

post #58 of 257
Quote:
Originally Posted by Spelaeus View Post


 

Actually, using the ABX plugin would only be important if you knew which track was which going into this. Since we're already prevented from knowing which track is lossy and which is lossless just by looking at the track, using the ABX plugin in this scenario would only add in confounding variables since you then not only have to figure out whether track 1 or track 2 is lossless, but also whether track X or track Y is the same as track A or track B which creates much more opportunity for you to trick yourself into hearing the same thing, audio memory being what it is. It would be an incredibly poor way to set up an experiment.



The point of the DBT is to tell if you can detect a difference regardless of whether you (sighted)  think A is better than B. The ABX comparator allows you to quickly compare A vs X , in terms of audio memory you are correct it is much better to use short segments and go back and forth rather than listening to whole tracks.

 

The problem with AB testing where you know that A is <> B is that you can imagine all sorts of differences. In the ABX DBT protocol the identity of X changes randomly so you are forced to do active comparisons such as is X = A or not, you do not know in advance if A <> X or A = X .

 

The whole ABX DBT thing is NOT about deciding which is which but finding out if any difference can be discerned, it is irrelevant whether you think A or B is lossless if you cannot tell them apart. If you can reliably tell A and B apart then you can worry about which you prefer...

post #59 of 257


 

Quote:
Originally Posted by nick_charles View Post





The point of the DBT is to tell if you can detect a difference regardless of whether you (sighted)  think A is better than B. The ABX comparator allows you to quickly compare A vs X , in terms of audio memory you are correct it is much better to use short segments and go back and forth rather than listening to whole tracks.

 

The problem with AB testing where you know that A is <> B is that you can imagine all sorts of differences. In the ABX DBT protocol the identity of X changes randomly so you are forced to do active comparisons such as is X = A or not, you do not know in advance if A <> X or A = X .

 

The whole ABX DBT thing is NOT about deciding which is which but finding out if any difference can be discerned, it is irrelevant whether you think A or B is lossless if you cannot tell them apart. If you can reliably tell A and B apart then you can worry about which you prefer...

I agree that there is the question of imagined differences when you know that A is A and B is B. But I think that issue is better dealt with by testing over a large number of subjects and perhaps having two different groups where the track assigned as A or B is swapped. Again, given the nature of audio memory, I'd have to say that the repeated measures experimental model used by the ABX plugin is a mistake here as opposed to a more cross-sectional method.

 

I'd be more willing to support the ABX plugin if it was using the subject's own music and tracks that they were intimately familiar with, since that better represents the actual listening conditions that we're questioning. But using unfamiliar music... I'll come out and say it, the ABX plugin becomes confusing. You have a tendency to convince yourself that the differences you notice are imagined when it's an unfamiliar track and when you try to do very detailed and intensive analysis (assuming there are no very obvious digital artifacts). The ABX plugin forces that sort of analysis, which is the confounding variable that I had mentioned.
 


Edited by Spelaeus - 7/6/10 at 3:15pm
post #60 of 257
Quote:
Originally Posted by Spelaeus View Post

 

I'd be more willing to support the ABX plugin if it was using the subject's own music and tracks that they were intimately familiar with, since that better represents the actual listening conditions that we're questioning. But using unfamiliar music... I'll come out and say it, the ABX plugin becomes confusing. You have a tendency to convince yourself that the differences you notice are imagined when it's an unfamiliar track and when you try to do very detailed and intensive analysis (assuming there are no very obvious digital artifacts). The ABX plugin forces that sort of analysis, which is the confounding variable that I had mentioned.
 



There is nothing stopping someone from creating their own test tracks of beloved music. You rip a CD track to wav encode a copy of it as mp3 then decode that copy back to wav you then have two wav files the same size which you can compare.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Double blind test 128Kbps vs lossless? I'll be amazed if you can tell much difference