They're AAC 128, right? From what I've read, that bitrate has mixed results for being transparent from the original CD. I'd like to know if a lot of people have successfully ABXed music they bought from iTMS from LAME-alt-preset-standard. I suppose I could buy a few songs I already have and ABX them from the mp3 myself, but it'd also be good to hear from others' experiences.

Typically, how discernable from LAME-aps are files from iTMS?

Maybe this will answer your question

I think there is a low pass @ 128kbs. I haven't abxed it but i think it is pretty obvious. For me i can't normal pick out artifacts but lowpass is really annoying it makes the music sound dead. But I have not ABXed it so my opinion isn't quite scientific. I should also note i was using the QT encoder not an actually downloaded itunes AAC file.

The lowpass for the Quicktime AAC encoder is 16khz, which is quite easy to abx (for me).

Thanks for the replies and the link. I'm probably missing something really obvious, but that listening test only seemed to prove that the Quicktime encoder is the best of the various AAC encoders, with an overall 4.39/5. I'm not entirely sure what that means, and the site that explains the testing procedure (there's a link to a readme.txt on another site) seems to be down now. Would a 5/5 or something really close have been completely transparent from the source WAV? How does the 1-5 scale measure transparency? Thanks for any more info you could give.

Well, it proves that QuickTime --the encoder that iTunes uses-- was, at that time, better than any other AAC encoder at that bitrate. In this other test you will find that MPC is the best codec at that bitrate, with AAC vey close. So you can conclude that QT 128k is among the very best you can get at that bitrate, nothing else. Whether 128k of AAC encoding is good enough for you is your choice. You should try making your own ABX tests against -aps to see how much you can really hear.


In this tests, people are asked to rate any given sample on a 1 to 5 scale (being 1 the worse and 5 the best). In this test, there is no reference WAV, the best score is just that: THE BEST SOUNDING. This obviously means that the highest score is the most transparent, but just exactly how much is not measured by a test like this.

This test may help you keep it in context. The closer to 5 the score, the more transparent the encode (more difficult to differentiate from uncompressed source).

In particular, note this part of the test so rjamorim doesnt increase his stress level at some point in the future :

edit: removed extraneous period, added qualifier to last statement
Also, Ataque has clarified with regard to the AAC encoder test

Ah, now I understand. Thanks for the info.