[r010159]

Hello!

I have been comparing the output of two different music players using ALAC files. I use DiffMaker to compare the two results. I then play the resulting difference file. I find if I turn up the gain on my amplifier enough I can hear the music that I used in the comparison. What does this mean? Shouldn't there be a null result, or a different result if the two files do not compare? I am thinking that this may be due to software jitter. But when I have the DiffMaker adjust for these timing and volume discrepancies, I still get the same result. I do align the two audio files very carefully in Audacity.

What gives here?

Thank you!

Bob Graham

 

[stv014]

Difference extraction requires very accurate synchronization of the samples (in terms of delay, level, and frequency response), and Audio DiffMaker might not have done a perfect job. Even with a small error, most of the difference will be the original signal at a lower level and possibly with some filtering (e.g. if the delay is not matched exactly, it will be mostly high frequencies).
 
Also, did you use the same sound card for both playback and recording ? If not, then there are synchronization issues due to the slightly different clock frequencies, which also slowly change over time by a small amount.

 

[r010159]

Difference extraction requires very accurate synchronization of the samples (in terms of delay, level, and frequency response), and Audio DiffMaker might not have done a perfect job. Even with a small error, most of the difference will be the original signal at a lower level and possibly with some filtering (e.g. if the delay is not matched exactly, it will be mostly high frequencies).

Also, did you use the same sound card for both playback and recording ? If not, then there are synchronization issues due to the slightly different clock frequencies, which also slowly change over time by a small amount.

I just used Audacity and Sunflower to record directly from each audio player and then compared the results. I am using ALAC audio files. I am trying to determine if the difference is due to a type of software jitter, where there are periods of alignment, or small actual differences with one of the players. I did very carefully align them wrt each other using a highly magnified time axis.

 

[stv014]

I just used Audacity and Sunflower to record directly from each audio player and then compared the results.

 
Yes, but you need an analog loopback for that. Was that done using the same device for both output and input ? If you mean you recorded the digital output of the players, that should not be different with bit perfect playback (not that I would expect a practically significant real difference in the analog output either).
 
If you still have the samples, perhaps it would be easier to find out what the problem is if you uploaded them and the original (with a maximum length of 30 seconds), and posted the links.