Maybe the "engineer" in question here hasn't really thought this thru. He doesn't see the elephant in the room: Radio stations screw up audio intentionally.
A good unprocessed AM will always sound better than the typical FM that is playing the squashed audio loudness game, as most seem to be doing. I have no research to quote, but I'm sure that dynamics always surpasses bandwidth when you measure perceived audio quality. Well, I'll make an exception for my old Mazda that wouldn't even pass the letter S on the AM band.
I wonder if most people who debate the merits of audio processing realize that our atmosphere is a compressor. Air absorbs violent oscillations more than weak ones. This has the effect of causing sounds to sound more and more compressed as they get further away. A perfect illustration is fireworks. If a large one detonates close by, the explosion's leading transient will audibly be much louder than its subsequent ambient environmental reverberation. But if a large firework detonates very far away, you may not even hear the leading transient -- just the sudden instantaneous onset of reverberation.
People don't often think about this consciously, but we do notice it subconsciously throughout our lives, and it does train our minds to associate compressed sounds with things that are large and powerful. After all, you never hear events with highly compressed dynamics in nature unless they're large enough
to be heard from a great distance in the first place.
This is why some people like heavy processing. When it is applied to wet and synthetic (rather than dry and acoustic) music -- i.e. music with lots of low amplitude ambiance sandwiched between all the loud sounds (reverb trails, harmonic coloration, etc.), all those low amplitude sounds are sucked up in volume, making that music sound "larger than life" by triggering the psychological association we have between low dynamic range sounds and distant hugeness. And that adds a sense of excitement and power to
that kind of music.
On the other hand, when you apply compression to music that is mixed dry and possibly is also acoustic (no guitar fuzz, no synths, etc.), there is no low level ambiance to suck up, and therefore, the music can't be made to sound bigger, more powerful, or "larger than life" by reducing its dynamic range. Instead, without that illusion of increased enormity, the ear is freed to hear that the dynamics processing
actually shrinks the apparent size of the music by making it sound flatter. It's even worse when a multi-band stage is trying to pull up low ambiance in each band but finds nothing to suck up -- in that case, all the ear has left to perceive is the multi-band stage constantly torquing the EQ of the music, which layers in a sensation of artificiality on top of the aforementioned flatness.
If only modern processors could songpart in addition to daypart. The heaviest processing could then be applied only to the songs it was capable of dramatizing and adding the most excitement to, while the most "purist" distortion-free peak limiting only-style presets would kick in for songs that demanded no audible dynamics manipulation.