"Dynamics" are straightforward to measure as linearity and distortion over different power levels. There's no reason to inherently constrain the measurement to use sines, it can use any (dynamic) test signal you want if you put enough elbow grease into it (which I have). For example, you can use diff maker and run it over time synched files. There are other sneakey methods as well.
From my experience designing many speakers, I think he may be tripping up on a few things:
- certain subjective colourations can enhance perception of dynamics. For example, non flat frequency response can easily give perception of false dynamics especially if the upper mid range is lumpy
- room reverb fills in the temporal gaps between dynamics. I think "horn dynamics" are as much a function of their tighter radiation pattern as any high sensitivity
- thermal compression occurs. Fielder and many others published quite a bit in JAES over the years on what it looks like, and ways to help minimize it. The design info is out there
I attend many (many) live classical shows and live classical rarely "sounds" dynamic in the same way a stereo might. At times, you don't even know the sound is as loud as it is until your ear starts distorting. Good, clean, accurate stereo systems provide this same experience. First time I truly experienced it in my home was when I deployed NC400s with 1200W power supply.