I completely agree with your premise and analysis. I have had many iterations of my systems over the years and have inadvertently moved too detailed, too lean and analytical... too revealing. As I have said elsewhere I finally spent over a decade researching the real thing... 7th row center seats at the symphony for a decade... other venues. I realize there are different camps of components and gradually upgraded my systems to reproduce the "real thing" instead of what I call "sound spectacular".
I look at details versus musicality a little differently. A good "real thing" system does not have less details over a "highly detailed system", but it keeps them in appropriate proportion. My system has all the details... they are not "over amplified". So you can still hear the violinist move her foot... but now, you have to stain to hear it... just as you would have to if you were in the music hall. It is getting the proportions right. I remember listening to White Room at a friends house once and being blown away that the cymbals were a solo instrument! They are not. I listened to the same recording on my system (carefully constructed to accurately reproduce music) and they were in the background by the drums where they should be.
So, it is a question of what are the manufactures trying to do. Companies like Audio Research, Conrad Johnson, Sonus Faber, VAC and a few others are dedicated to the accurate whole reproduction of music (in 3D space)... with all instruments in proportion to what is on the recording... and would be in real life. Then there are companies that carefully reproduce each sound in 3d space emphasizing all the details and revealing all. These can be amazing to listen to... but do not sound like you are sitting where the microphones were placed... in my example, in the symphony. I sat about three rows back and 12' below the microphones used to record the symphony when it played and have those recordings... some of which I was at during recording.

