[brookst]

A few years ago some friends and I did a carefully controlled blind test between 320kbps mp3 and FLAC. Studio monitors, identical source HW, etc.

The person playing them was not in the test, we randomized 20 second samples of different styles, randomized order determined in advance.

I don’t remember exact results but it was completely inline with binomial probability.

To this day I feel like there’s something wrong with high frequencies like open hats in 320 mp3, but on that day I could not tell a difference so I’m accepting it as negative placebo.