I carry "memories" of live music, and still attend performances from time to time (but recently I hardly ever get the time) but I tend not to use live music as a benchmark except for some classical music - specifically chamber music. For me, there are very few times when I have been able to get the seat at the venue I would have needed to get in order to achieve the sound I can get at home. In addition, a lot of performances are amplified and the music sent out over speakers and the music competes with the girl next to me shouting to her boyfriend how TOOOOHHHHTALLY AWESOME the whole thing is, the guy in front puking his guts out, and the person behind me spilling beer he smuggled in. So much for rock concerts.
Jazz is OK, but jazz and classical concerts for me are so much more about appreciating an interpretation of music you already know. There are, as Detlof knows, umpteen versions of the Goldberg variations, and I'd swear I can hum along with the best of them and not get a note wrong, but I can still be surprised by little things when I hear a performance, and it is the interpretation, NOT the sound, that I seek when I go out of my way to listen to live music. And THAT benchmark is tough to emulate at home.
My personal 'benchmark' for my system is more an 'image' of what that particular performance SHOULD sound like. It may not be a memory per se; instead it may be an image created in my head, a mishmash (partly based on memory) of knowing what a given musician should be able to do with a given piece of music and a given group of people, along with the instruments, the temperature, whether it was the first take or the 5th, and trying to figure out whether the conductor or ensemble really was satisfied with what they did. Indeed, the chance to actually compare live with 'memorex' is indeed very small, but that doesn't bother me in the slightest - the live performances I appreciate the most are the ones which would lose their charm if repeated anyway.