An age old controversy: The objectivists vs. the subjectivists
With an engineering background myself, I started out firmly in the objective camp, and assumed we could 'measure our way out' of anything. But over the years Ive found my personal truth lies somewhere between abstract measurements and the gossamer veil of subjectivity.
The big key for me: We dont know exactly how our brain processes the aural stimuli. Oh, we know all about the mechanics of it, but far from how it is perceived once the stimuli enters the brain.
Because of this fundamental shortcoming, we dont even know how to measure what we hear, or the relative importance of the measurements we do make. We only know what we hear sounds good or bad. (Score one for the subjectivists)
For example, it is very easy to measure phase distortion, but even severe phase distortion (consider a 4th order crossover here, for instance) seems of little consequence as it appears our brains have no problem piecing the various parts back together into a cohesive whole. -Likely a defense mechanism trait learned eons ago in localizing our predators.
On the other hand, even tiny amounts of high order harmonic distortion or inter-modulation distortion can be painfully obvious, but even our most elaborate test equipment may have issues pulling them out of the inherent noise floor.
Consider two full range drivers, both with ruler flat frequency response and harmonic distortion figures down in the low tenths of a percent or -60 to -70 dB down from the fundamental. One would expect them to have a similar sonic signature, or perhaps no sonic signature at all, yet they may sound different due to well, we just dont know. IM, Doppler, SIM, TIM, or some combination? Your guess is as good as mine...
Resolution:
Those familiar with distortion plots know there is no Zero, but a noise floor that is inherent in the device under test and also in the test equipment itself. In audio, I suggest this exhibits itself as a white noise background that our brains tend to process out, (but I suspect is one source of listener fatigue.) If the source material contains musical information that approaches the level of the noise floor, this information is masked by the noise. An example of this might be a singer taking a breath before starting to sing. In a system capable of good resolution one may just hear the breath. In a mid-fi system, you may not, as the information is masked by the higher noise floor. Note the difference in noise floor may not even be audible, and conventional tests of THD+N may not show any difference in level between the devices.
Until we learn how the brain processes what we hear, I suggest we have to depend on our ears as well as our test equipment. Neither is sufficient in itself.
C
