I don't listen a lot with headphones, but I suspect what's missing from the headphone experience is the interaction between the loudspeakers and the room. I've listened to so much live jazz that I have the characteristics of that live sound indelibly burned in my brain, and my "holy grail" reference for evaluating sound on my 2-channel system is to close my eyes and evaluate how close jazz sounds like it's being performed live in that room. That "live in the room" evaluation doesn't work at all for rock (for me at least), but it works really well for jazz. The closer I get to being happy with the system, the more it sounds like the speakers are interacting with the room in the exact same way the sound from the musicians' sound would interact if they were sprawled across the stage from left to right speaker. It's kind of difficult to explain, and maybe completely nonsensical to some since you want typically want to reduce room interactions -- but musicians playing in the room would absolutely interact with the room reflections and I feel speakers must interact in the same way to recreate that "live" feel.
I guess the short version of this is that listening to headphones is more like listening to a performance in the studio, and listening on speakers is more like a live performance with the musicians sitting in your room with you. I guess with room treatments, you could probably narrow the differences loudspeaker and headphone sound. That would certainly measure better and may be the technically better solution -- but I'm not sure it would help my personal preference of having it sound like it's live in my individual room.
