Hans,

Interesting setup you've got. How do you do a double-blind test to remove the placebo effect when comparing flac vs. mp3? Do you guys do both perceptional tests as well as pure analytical comparisions?

Analytically, no question mp3 is going to be different from flac or the original source, and it's going to be a lot easier to quantify those differences. And since mp3 (and lossy compression in general) is about perception, it's going to fail pure analytical comparisions.

Almost everyone can't tell the difference between original souce and mp3's encoded by lame using the --preset standard settings. There =are= a few people who have repeatedly demonstrated the ability to distinguish lossy encoded files.

But all of those test setups have only been done via software on pcs, and not on very high-end equipment. And the few times I've been able to listen to music on crazy high-end gear, I do believe it makes a difference.

I wonder if reliable and repeatable ABX testing on very high end audio equipment would increase the numbers of people who distinguish compressed music from the original, or change the qualtity level that it can be distinguished at?

--Nathan