The inclusion of a compressor or wavehammer-like plugin of the DSP Studio. GComp works about 70% satisfactorily. I live in an MDU and it's necessary because perceived volume cannot jump, end of, else I get complaints from the neighbours, whatever the gain is. The technical document in the wiki states specifically that the process within MC uses peak level, but peak level is not an indicator of loudness. Root mean square (i.e. basically the mean of the difference between the peaks and valleys in the waveform) is, so I find it fundamentally inadequate.
Ex, a typical minimal dance track ripped from vinyl you can have peak at or near 0, but its RMS is going to be around -18db to -15db as opposed to a radio show, compressed during broadcast, also, peaking at 0, that will have it around -10db or even higher. They would be at the same peak, but the difference in RMS is going to make the radio show seem much louder. -10db I find to be the practical "listenable" loudness, but some loudness war era things have it much higher (i.e. there's spots on that "Iggy Mix" of Iggy Pop's Raw Power that reach -4db).
I want the operation to occur on the fly as part of the output chain, rather than changing the files themselves. Sony's Wavehammer used to work well for this but the gui broke around MC15.