You seem to assume that we know exactly which objective parameter correlates with the subjective perceptual effect in question.
No, what I'm assuming is that if the perceptual result is harshness (distortion) in a specific range - upper midrange and treble - that it will show up as non-linear behavior in the frequency response. There may be even better ways to measure it, I'm not an engineer. I also assume that good DAC developers, especially on the pro side,
are measuring and isolating distortions all the time, so they can find them and reduce or eliminate them. Hell, I'd further assume that they
have tracked them down and know that these days, most "glare" comes from recordings, rooms and speakers.
Hopefully, you are not a strict objectivist, of the kind who declare that if traditional objective measures appear to reveal nothing of consequence then any contrary subjective perceptions are necessarily false.
I don't go quite that far, no. But in this case, what people are describing perceptually shouldn't be that hard to measure objectively. If we were talking about sound stage or PRAT, it would be a very different conversation.
A problem I have with traditional audio measurements is that the test signals utilized are usually static, while music content is dynamic.
I don't think that's as much of a problem as many people do. A lot of excellent equipment has been designed using such test signals, but I would prefer measuring the actual output of a system, in an anechoic chamber. Of course then we'd have to argue over the quality of the amplification, monitors and cables. That doesn't concern me much; unlike many audiophiles, I have great confidence that there are pro monitoring systems that would easily reveal any audible anomalies in the upper mids and treble, both in the ears and in the numbers.
Further, should we come to understand the relevant physical parameter causing some perceptual effect, we would also need to know the sensitivity of that parameter with respect to with the human pshycho-acoustic system in order to recognize it's significance. Glare may be the result of some complex distortion intermodulation spectra, or dynamic changes in harmonic distortion spectra profile, or low level quantization noise floor modulation, or jitter correlated distortion/noise modulation, or time-domain distortion due to insufficiently sharp band-limiting of the signal, or god knows what else.
I think I just had a little acid flashback...
Tim