Golden Ears and MP3

I was cruising around one of the audio forums and came across a discussion about pre-amps and how the RNP from FMR audio stacked up against several other brands. I use the RNP. I used to use an Audio-Buddy from M-Audio. The RNP seems nice, but I don't remember enough about the sound of the Audio Buddy to say for sure that I could tell the difference between my old pre and this new one. There are so many other things that have a bigger effect - strings, guitar, microphone, microphone position; don't forget ability - which I hope is changing for the better. Our auditory memory is pretty short also - maybe a few minutes, if that. So I was kind of interested to hear comparisons and opinions of experts.
One of the participants created a blind 'shoot-out', just him and his guitar going into 4 different pre's. There were a couple of things that rubbed me wrong - he clearly had an agenda and he was the one making the guitar recordings and I believe some of his agenda crept into the recordings. So number one, it wasn't double blind. But the most egregious thing he did was post the results using 128 bit mp3. And this isn't the first time I've seen someone go to a lot of trouble doing sound comparisons only to post the results in some lossy, frequency-limited format.

Mp3 uses some tricks to encode files, one of which is to leave out information. The lower the bit rate, the more information is dropped. An mp3 file can take up 10% or less of the original space used for the file, but beware the trade-offs. There are only a few ways to compress music files so that no information is lost. FLAC is the one I use to store wav files, but there are a few others. Flac compresses a file about 50%.

The waveform at the top of the article is a wav file seven seconds long, recorded at 44.1 khz, 16 bit. This is how a CD sounds. I encoded this wav to an MP3 at 128 bits, then I inverted the waveform and subtracted it from the original. What remains is shown here:
This is the frequency spectrum of sound that didn't make it into the mp3. Not convinced, yet? You can hear the sound that was left out here.

I can hear differences between pre's and microphones in 320 kbit mp3 files like the ones posted at The Listening Sessions. Anything compressed more than that is a guessing game.

Update:
Another reason not to use mp3 encoding:
Microsoft Loses Big In MP3 Patent Suit - washingtonpost.com
I think some spectrum charts of Ogg Vorbis might be interesting.