Old thread, but I got pulled in and it relates to my current plight.
Clearly, there is no simple answer, but lots of good ideas here.
I have recently been helping an FM Rock radio station tweak their processing. I'm listening to a budget 2005 Technics receiver with vintage Hegemen Model 1A omni-directional, top firing speakers. They sound pretty darned good, especially considering they are small (8" main speaker with a modest tweeter), I have no sub, no EQ, no "loudness" button, just a basic, pleasant system in my room to watch TV and listen to music on.
Wednesday the station played an acoustic Dylan song and I felt a rumble that resembled acoustic feed back from a turntable, though the source file was ripped from CD and played from a computer. Perhaps it was slight HVAC noise, recorded, but not noticed during Dylan's original session, being expanded by the signal processing. Today, during a live, acoustic Neil Young song I felt a loud thump on every beat. I put a high-res spectrum analyzer on my tuner's output and noticed a huge peak at 36 Hz, so it was probably on the original recording, just not evident without massive compression or expansion. Could be some kind of "Max-bass" type of harmonic enhancer at the station, but that seems less likely.
Now, how are we to anticipate what our music will sound like under those circumstances? Getting a good mix on good equipment is hard enough. Bit rate reduction, radio processing and other sound mangling process are too numerous to contemplate and there are new algorithms being written daily.
Conceptually, it would seem prudent to filter sub-sonic energy. Inaudible audio you leave on your master could get quite nasty in the real world, as I learned this week. But where is the line;20 Hz, 30, 50? High pass, band pass, both? How steep the slope of your filter, how wide the Q? This alone gets dicey. High pass filters create a very small bump at the cutoff frequency. Put one on every track set to the same frequency and you will end up with a big hump when the channels are summed. So it's usually better to run a bunch of tracks you plan to filter in the same way to an aux and use one EQ to avoid this problem, that is, if yon need to roll off the lows.
Before you release your master, play it back many times and see if it can stand up to the tortures it could be subjected to in the world. Squash the hell out it with compressors, multi-band if possible. Fiddle with AGC, gating, upward expansion, EQ, make low bit-rate MP3s, etc. Once it leaves your hands, it's anybody's guess what will happen to it. If you can spot potential dangers and prevent them without doing violence to your mix, you will be a step ahead of your average producer, maybe even ahead of some legendary names.
I don't have an answer, but after 46 years working with audio I discovered something new and disturbing this week. Call it food for thought.