[RonaldDumsfeld]

Well in that case just post them up and see if anyone else can do it.

 

[StanD]

  Well in that case just post them up and see if anyone else can do it.

+1
Yeah, like Rummy sez.

 

[adupree]

  There are also limits to what human beings can perceive as we are not test equipment. Did you null balance the two files? I suspect that you didn't.

Correct some things can't be heard, but people can hear very small differences. In the music industry mastering engineers are perfect examples of this. 
 
If the audio level is the same the inversion of the wave would cancel. 

 

[adupree]

  Well in that case just post them up and see if anyone else can do it.

I'll be happy to, if you can tell me how to import a file.

 

[RonaldDumsfeld]

I don't think you can do it directly on this site. Perhaps someone more experienced can help. Or filesharing site?
 
The best place to this kind of thing is over on Hydrogen Audio.
 
http://www.hydrogenaudio.org/forums/index.php?act=idx
 
They have a special forum for uploads. If there is anyone who can replicate and confirm your results that's were you will find them. Would be cool if you have found a problem sample because it will help improve the encoding in future releases.There was a guy who used to post here (UltMusicSnob) who could detect differences between WAV and MP3 in a piece of classical music with a very wide dynamic range.
 
So it can be done. Give it a go mate. You could become famous. Or at least a member of an exclusive club.

 

[bigshot]

  When you listen to the different files they do not sound the same. It's as simple as that. Top-end and some low-end are missing in the lossy files.

 
Take a look at the spectrum in an audio editor. At 320 AAC most of the difference is above 18kHz where there isn't much music anyway.

 

[bigshot]

  There was a guy who used to post here (UltMusicSnob) who could detect differences between WAV and MP3 in a piece of classical music with a very wide dynamic range.

 
He SAID he could detect differences and posted up copy/paste text that SUGGESTED he could. No one ever verified it.

 

[adupree]

   
Take a look at the spectrum in an audio editor. At 320 AAC most of the difference is above 18kHz where there isn't much music anyway.

Correct, but there still is a difference. Think of it like this, mastering engineers make small adjustments, sometimes a .5dB cut or boost somewhere. Maybe very very light compression somewhere, but it can be heard. So how would this be different?

 

[bigshot]

Is that something like one or two less angels dancing on the head of a pin?
 
A half decibel on one channel in a full mix? I don't think what you are talking about is audible. And I know it really doesn't make a lick of difference to the quality of the sound.

 

[mikeaj]

Furthermore, the "difference signal" from lossy compression is generally engineered to be masked by the actual music (or rather, the algorithm is designed that way; it may not produce best results 100% of the time but a well-tuned one should do well generally) in a way a certain boost or cut or other change made during mastering wouldn't be.
 
Anyway, a difference is a difference. We're just talking about the magnitude and characteristics of it.

 

[bigshot]

audibility

 

[mikeaj]

yup, something small enough and/or with the right characteristic should not be audible

 

[adupree]

Then guys like Bob Katz and Bernie Grundman must be insane for being so precise. But hey what do I know. No sense in debating anymore.

 

[StanD]

  Correct some things can't be heard, but people can hear very small differences. In the music industry mastering engineers are perfect examples of this. 
 
If the audio level is the same the inversion of the wave would cancel. 

The overall level doesn't have to be the same in two files that have been processed differently, you must null balance them to determine the actual differences. Otherwise your test is useless.

 

[mikeaj]

I personally didn't define or sign onto any precise description of "small enough" 

 
I'd say the substantive changes in presentation are generally more than slight 0.5 dB cuts or boosts, but things do add up... and so on. And there's nothing wrong with doing a good job. And let's not keep going in order to drag discussions down into the stupid or strawman range. 
 
 
With respect to the files themselves and compression, please refer to RonaldDumsfeld's response. Legitimate positive results aren't unheard of. Sometimes systems don't work as intended. Also see above discussion with respect to nulling.