Scientific methods can be used to measure all sorts of performance criteria- volts, amps, volume, etc. Human beings however listen and rate performance based on their own personal evaluations of all of these combined. Scientific methods can be used to try and determine things like the weighting a particular listener gives each measurement, but it would be an inference based on their actual choices. This is all a long-winded explanation because you said, "science cannot predict how humans perceive sound" which is correct. But you then finish the sentence with, "but should be able to measure the performance of equipment!"
But science can. The discipline you seek is called logic or deductive reasoning. In it you will learn to avoid combining unrelated clauses as if they lead to a conclusion. They don't. They are unrelated. They lead to naught.