Hey! No psychoacoustic model can be marked as perfect, but let's do a try!
I'd like a utility that runs MPC's quality 10 psychoacoustic model' parameters as like a digital pretty good pseudo-hearing, then you give two samples to it: original and wav from a encoded source (with any encoder at any bitrate) then, apply various criteria to see both kaps (kiloartifacts/sec) and NSR (noise to sound ratio).

That will show people using Xing encoders how horrible are their encodes@320kbps!!

C'mon! y' have only 340'42Maps and NSR of 210dB! The most pure sound ever seen! Almost lossless!!!