WAV or Apple Lossless Encoder?

I think everyone should try this at home and report the results... I was alerted to this by a leading manufacturer of PC audio related gear (who can identify himself), and found it to be true. Here is the test:

1) Burn a song from a clean CD to a wav file using EAC (this is now very easy to do, the software is free, downloadable, and very user friendly).

2) Import that wav file into Itunes and convert it to an apple lossless file (you need your default import setting set to lossless to do this).

3) Using the same CD, rip the same song directly into Itunes in lossless format using error correction.

4) Compare the two lossless files.

I've done this, and I was astonishly saddened to hear the results (because I ripped all my music in lossless). The lossless file generated through Itunes had less seperation b/t instruments (poorer imaging), sounded slightly less "alive", and (I believe) more "smearing" of the sound. I haven't done this yet, but I wonder if the same observation would be made through an Itunes generated WAV file vs. an Itunes generated lossless file.

I'm convinced this is real. I asked a non-audiophile friend to do the same listening comparison, told her nothing about what she was hearing, and she made the same observations. Marco (Jax2) will do the test tonight.

WE DO NEED DATA, AND I THINK THE MOST CREDIBLE SOURCE IS MULTIPLE PEOPLE DOING THIS COMPARISON. PLEASE DO YOURS TODAY :) - AND REPORT BACK WHAT YOU FIND.

Peter, You have good ears. I have been ripping to Apple Lossless for a few years and I always thought they sounded the same as WAV or AIFF. But a few months ago I experimented something similar to what you did. I ripped some CDs to AIFF using iTunes, converted them to Apple Lossless, burned two sets of CDs: one directly from AIFF, another one directly from Apple Lossless. I than played them back from CD player. Some tracks from the Apple Lossless ones sounded thin and some with narrower soundstage and not as dynamic. It didn't happen on all the tracks but the ones did sound different were very obvious, even my non-audiophile friends can point that out immediately.

The Apple Lossless format didn't loss any information because when I converted them back to AIFF and compare them to the original AIFF files, they were identical. That lead me to believe that the one-step conversion process during CD burning may have introduced artifacts.

Interesting observations.

Sidssp - Did you try ripping the tracks that sounded different on the burned CDs and comparing them to each other and to the original ripped tracks? It would be interesting to see if the bits are different. If so, then the burning process may be the issue - which I think is what I think you concluded.

Peter_S What do you think caused the differences you hear? Do you agree if the bits are the same, the files should sound the same? If so, you might want to compare the actual files bit by bit. It is certainly possible that there is some error in the conversion process. You could compare the Apple lossless files with a file compare utility like fc/b, although it is pretty crude. You can also convert the Apple lossless to wav and compare those files using EAC or foobar. As I said above, when I ripped tracks to Apple lossless with iTunes and converted them to wav and compared to EAC ripped tracks, they were identical. Have not tried a bit compare after doing a wav to Apple lossless conversion but could try it. Also you could convert the Apple lossless lossless file that was converted from the EAC wav file back to wav and compare it to the original EAC wav file to see if the round trip introduced differences. EAC has a good comparison routine, although you have to set the disk offsets to 0 when ripping with EAC to do the comparison. Another interesting thing to do is to take the EAC wav and a iTunes ripped Apple lossless file converted to wav and use the ABX facility in foobar to do a blind test of the two files. Much harder to do comparisons with iTunes. (I don't want to get into a discussion of blinding testing - just wanted to point out the utility in foobar.)

Peter dropped off several files for me to listen to this evening. He didn't tell me much about them. I didn't have my server running at the time and our dogs were playing so we could not listen to them together, so I don't know if my observations are the same as his. I spent about thirty minutes going back and forth between various versions of a rip of Rickie Lee Jones, Bye Bye Blackbird off of "PopPop". I had one of my own which was a WAV ripped in iTunes using Error Correction. Peter provided three files. I could only locate two of the three as the third evidently did not have some critical tags and got sent into oblivion by SlimServer. So of the three files I had, one that he provided sounded significantly different. It seemed to have an improved soundstage with better imaging, more natural vocal presence, and deeper bass. The WAV I'd ripped via iTunes, and the other file he provided sounded quite similar. They did not provide as well defined a soundstage as the one file. I don't know if he was trying to trick me, but I phoned him with the file that stood out to my ears and in my system - he can confirm whether or not my observations paralleled his, or were different. There is no doubt at all that the files he provided sounded different from each other, and that one sounded better than my iTunes WAV rip of the same file. His version of PopPop is the same as mine, but obviously it is a different physical disc, ripped on a different computer.

OK, Peter, was the file that sounded better to me the same as the one that you preferred?

Peter is not the only one that has heard this anomaly by a long shot. I have received a number of reports of the same thing. The problem happens even when the data is streamed with WiFi to AirPort Express or AppleTV.

If the file is ripped directly to Apple Lossless with iTunes, the quality is inferior to the same file ripped with EAC and then converted to Apple Lossless with iTunes.

If the ALAC data is converted back to .wav, the files compare, but this does not tell the whole story. For instance, one file might be 24-bit and the other 16-bit, which makes a big difference in quality. They will still compare.

Steve N.