[nick_charles]

One of these files is a bit-perfect rip of a short CD track. One is a digital recording via an Edirol UA-1EX from the optical output of a Marantz CC4300 but is **verifiably** not bit-perfect. The other is an AD recording via an Edirol UA-1EX from the analog output from a Marantz CC4300. I was unable to correctly DBT them but hopefully younger ears can do better. All files are bog-standard 16/44.1 wav files with no DSP or resampling. Minimum trials 10 please

 
X.wav
 
Y.wav
 
Z.wav
 
Hmm, I was able to find a segment at about 1;09 where I could successfully DBT the analog recording and the reference file 13/15
 
Damn Damn Damn ! - there is distortion at 1:09 (on all copies) I had never heard it before, now I'll never not hear it again damn !
EDIT: New versions, silence trimmed , randomly renamed...  
A.wav
 
B.wav
 
C.wav

 

[Anaxilus]

Do you know there are errors between the samples or just assume that not using bit matching produced errors?  Kind of irrelevant anyway since I would go bit matched for archival purposes.  Might as well.

 

[nick_charles]

Quote:

Do you know there are errors between the samples or just assume that not using bit matching produced errors?  Kind of irrelevant anyway since I would go bit matched for archival purposes.  Might as well.

Yes, the three samples *are* measurably different. It is just an exercise and yes I would never actually do an A/D recording of a CD for listening purposes that would be plain daft. I just wanted to explore just how audible the differences might be as some folks here are a bit OCD about having bit-perfection

 

[Anaxilus]

 

 

[Ham Sandwich]

How would you go about doing a three-way DBT like that?  You can DBT two files against each other in Foobar.  How do you do three?

 

[nick_charles]

Quote:

How would you go about doing a three-way DBT like that?  You can DBT two files against each other in Foobar.  How do you do three?

You don't. You do 3 DBTs AB AC BC.
 
But lets keep it simple. X is the Analog recording and detectably different by my 51 year old ears so not a serious test it turns out after all.
 
So the other two are a perfect rip and an imperfect copy.
 

 

[xnor]

that download site sends me into an endless loop, or maybe I just dont see the link between all the ads

 

[nick_charles]

Quote:

that download site sends me into an endless loop, or maybe I just dont see the link between all the ads

Cick Download button , wait the required 15 seconds then click Download button again

 

[xnor]

I did just that and it didn't work, but guess what, it mystically works now. Weird...
 
edit: for the first file, at least hehe
Maybe there's a max concurrent downloads (1) limit or something.

 

[nick_charles]

Quote:

I did just that and it didn't work, but guess what, it mystically works now. Weird...
 
edit: for the first file, at least hehe
Maybe there's a max concurrent downloads (1) limit or something.

Patience is a virtue, not just a card game

 

[Pio2001]

Hello,
I couldn't ABX X vs Y at normal listening levels with Sennheiser HD600 headphones. However, statistics says that the RMS level is 0.5 dB quieter in the right channel of X, while the left channels are quite the same. The ABX should thus be possible relying on the balance, that is a little to the left on the X file.
The difference between Y and Z is just flat dither, +/- 1 level. Its RMS level is -96 dB.
 
It is therefore possible to cheat and ABX all files cranking up the volume to insane levels, since there is a silence at the beginning.
 
Results with Superex Pro-B VI closed headphones, volume cranked up 26 dB above normal listening level, listening to the initial silent part only :
X vs Y : ABX 10/10. X has analog hiss.
Y vs Z : ABX 10/10. Y has permanent white background noise while Z begins with a fade-in.

 

[nick_charles]

Hello,
I couldn't ABX X vs Y at normal listening levels with Sennheiser HD600 headphones. However, statistics says that the RMS level is 0.5 dB quieter in the right channel of X, while the left channels are quite the same. The ABX should thus be possible relying on the balance, that is a little to the left on the X file.
The difference between Y and Z is just flat dither, +/- 1 level. Its RMS level is -96 dB.
 
It is therefore possible to cheat and ABX all files cranking up the volume to insane levels, since there is a silence at the beginning.
 
Results with Superex Pro-B VI closed headphones, volume cranked up 26 dB above normal listening level, listening to the initial silent part only :
X vs Y : ABX 10/10. X has analog hiss.
Y vs Z : ABX 10/10. Y has permanent white background noise while Z begins with a fade-in.

New versions mounted (A,B,C) randomly renamed with leading silence trimmed.  
A.wav
 
B.wav
 
C.wav

 

[Pio2001]

Can't tell the difference. I didn't try to cheat with the ending silence.

 

[Pio2001]

I just realize that I had done yesterday's test (X/Y/Z) with Kernel Streaming drivers, while I did today's with DirectSound drivers, that are not bit-perfect

 

[fishski13]

no random DBT here, but A and C sound identical, with B different. 
 
good one N.C., and thanks for putting time into this!