@amirm thanks for the video links! after reading several different threads, it's great to get the basics.
I have too many questions now...
I understand that frequency response is the single most important factor in speaker preference, and that mostly people prefer the same speakers, whether they're ordinary people or well trained. Great. That makes sense.
People adapt to their speakers, as you say. Our brains can correct the frequency response and spinorama results, to some extent, maybe better than equalization can. I bet we can even correct or get used to small dips that you can't easily fix with DSP. I wonder: What's left after people's ears and brains have adapted to a particular speaker? What can't our brains correct?
Distortion and noise cause information loss. Our brains can't recover information that isn't there. Speakers with a perfect spinorama with high distortion might have less musical information than speakers with jagged spinoramas and very low distortion. Our brains might EQ the latter speakers and prefer them after the "break-in" period where our brains rewire.
What other information loss could occur? There must be something happening in crossover regions, where lobing and imperfectly aligned sound waves intersect. Is there? Does that cause information loss that our brains can't learn to put back? Does the FR and spinorama capture that? Maybe speakers with worse FR would be subjectively preferable, if the tweeters were AMT or beryllium or ribbons, or the crossovers were concentric or time and phase aligned.
I guess I'm wondering if speaker preference also incorporates a little information theory. An imperfection that doesn't cause information loss can be corrected by our brains. An imperfection that causes information loss cannot. After we rewire our brains a little bit from extended listening, we can correct all the non-lossy deficiencies in the speakers. Of course, it's better not to have to do this. The most important factor in subjective preference will be the listener's initial perception of things like FR. But maybe there are other things that matter too, from an information-theoretic perspective.
To come back to the topic of this thread, what might there be about ribbons or AMT tweeters that might make a subjective difference and not show up in the standard plots that are shown on this site? Anything?
Full disclosure: I've owned speakers with aluminum dome tweeters, soft dome, AMT, and concentric drivers. I've always felt there was more to the difference than what is captured in the FR, distortion, and off-axis FR. After spending some time on ASR, now idk!