Tim, the cues used in audio perception are in our nature - our auditory perception tell us what sounds belong to what physical objects & does a pretty good job of it - it is part of the necessary survival role that has evolved our perceptions. Your question does bring us back to my first post. Would you say that somebody developing an audio product by measurement alone is correct or mistaken in this approach?
Well my argument is that auditory perception has it's own criteria for judging how unreal a reproduction is & we are only beginning to ascertain what these criteria might be. As per the Head movement paper, this researched aspect is at a very low level (microseconds timing differences &/or decimal DB amplitude difference). If we don't know the criteria & level that our perception expects then yes, we can probably end up with an audio device that measures great as per current measurement sets but perceptually sounds bad. I wouldn't exaggerate this & say "gotten all these perceptual cues terribly wrong?" - we don't know what are the important cues as far as perception is concerned & we don't know what the JND is for these cues.
I don't buy the argument "anything that we can perceive will show up in the low level measurements" for the reasons I've already given - because no model of auditory perception is being used & therefore the wrong test signals & wrong analysis are the result. As I said, since the discovery of DNA's molecular structure through xray diffraction, we haven't continued to use xray diffraction to try to analyse the differences between DNA between different species - we haven't used electron microscopy either - we have used a technique that is appropriate to uncovering the nucleic acid sequences (which, btw only gets us a small bit of the way towards understanding how biological objects are formed from the underlying DNA structure).
