I've used RTs in several songs and always rendered individual instruments to a WAV file and then transferred them to my hard disc recorder for mixing. Never really noticed the problem described here. I think I'll record the audio out of my laptop and compare it to a rendered WAV, after normalizing to the same peak dB.

If there is a significant difference, then this is a serious bug in rendering to WAVs.