My Dad (a retired electrical engineer) is neither a tone hound nor an audiophile but he's often remarked to me how, when solid state stereos came out, people like him were so wowed by the specs of the SS gear than they failed to notice that it didn't sound as good as the tube stuff.
There's more to the story than that. Some - not all - early ss stereo amps had crossover distortion, which was the result of poor engineering work, not an intrinsic characteristic of transistors. This type of distortion is at a maximum
percentage at the lowest signal levels, and it is obnoxious-sounding. The "specsmanship" that covered up this onerous behavior was to disclose distortion at
maximum power only. Had the disclosure included distortion at small-signal levels (a watt or less), the distortion figure would have been so high that nobody would have ever considered buying those POS products.
It is the unfortunate lapse in the work done by the EEs who designed some of these amplifiers (and by the marketers of the products) that led directly to the myth that "tube amps sound better than ss ones." This is a textbook example of generalizing falsely from one or two specific instances, and it is the same kind of flawed "reasoning" that leads to "digital sounds harsh," and all sorts of other audio myths. As I am fond of pointing out, it is one thing to observe an
effect - almost anyone can do that - but another thing altogether to correctly identify the
cause, which most often requires insights that go far beyond those that can be acquired by reading enthusiast publications and manufacturers' ads.
There is no formula or specification for what sounds good,
While this is true, it misses the point of using FRFR. You're not
looking for "sounds good," you're looking for "sounds exactly like what you put into it." If what you put
in "sounds good," then a proper FRFR system will
preserve that "goodness," neither adding anything nor taking anything away. There
is a "formula," albeit a very complex one, for determining how well a given loudspeaker meets that criterion. And, given the opportunity to compare loudspeakers, it
is possible to train your ear-brain system to identify the one which does the best job of preserving the signal at its input.
particularly when the spec comes from a manufacturer playing a specs game.
That will continue to be a problem. When an entire industry is in the habit of "enhancing" product data - as is the audio industry - then the "data" that is available to prospective buyers of a product is useless. That is
not proof that you can't identify sonic attributes with data, however. It is most definitely possible to acquire data that correlates directly with sonic performance. Unfortunately, it won't ever be an option for the buying public.
We need to go with what sounds good to our ears, not our eyes.
Ears are the final test instrument. As I have pointed out many times in the past, the evaluation process in which the ears are involved, if it is to be successful, must be disciplined and thorough. Going with what sounds "best" in a poorly-arranged, brief "shootout" will often as not lead to disappointment.