Not sure if these have been posted before;

thought they might be good for someone;
Actually stolen from a post at StorageReview

http://www.geocities.com/altbinariessounds...al/mp3test.html

http://www.heise.de/newsticker/meldung/30571
This one is in german about comparing various codecs

Isn't that the test testing only one sample, and an artificial sample on top of that?

seems like ogg vorbis is the winner of this test (best at 64kbit and no big differences at 128k and above between the codecs). also vorbis has perfect tagging, gapless playback, replaygain and it's open source. i think it's likely that vorbis wins this test (overall).

(i'm refering to the german heise article)

They did test only one sample, which was a collage of 3 (or 4?) very short clips. Not really artificial, but far from representative, of course.
Also, please note the date on top of the articles. One is 18 months old, the other almost 4 years, so the results are definitely outdated.

edit: Although this test lacks the professionalism of Roberto's public listening tests (not only the sample was flawed, but the testing methodology as well), they had at least one thing to be envied for: 6000 participants.

Hehe. Talk about their methodology