I'm just happy I came across that article demonstrating the fallacy of believing static measurements is all there is to know, and needed.
Three things:
1. The stimulus is not static. A sine wave is time varying. A multitone is time varying and complex. Jitter signal (J-test) is also complex and time varying. Only DC signals are static which we don't use for audio testing.
2. For the most part, audio gear is state and memory less. The system performance doesn't change or rely on what came 1 second before current time. In that regard, "static" testing of this sort is quite appropriate.
When testing systems that do have memory such as lossy audio compression (e.g. MP3), we cannot use this type of testing because they do have memory and adapt to signal to being encoded. There, we rely on controlled/blind listening tests.
3. I also using music files for such things as null tests of power cables and such. This is more done to do away with objections like yours more than being a need.