An old friend visited recently -- a close friend and a music lover who lives in Virgina and whom we rarely see any more. As he examined our living room with the home theater set up at one end and the hi-fi at the other, he asked "Why do you have all these speakers around the TV but only two set up for music? I thought surround sound was the next big thing?"

Well who wouldn't think that from reading the experts and the pundits? The fact is, I don't know anybody currently enjoying a two-channel system who is dying to have surround sound. But that last sentence, while completely accurate, is very cleverly constructed. What about all the people who aren't "enjoying" their stereos? They're the ones who want surround sound for music and they fall into two categories.

Some are simply folks like my erstwhile colleague J. Gordon Holt -- people who have always felt that stereo was an unsatisfying attempt to portray the sound of music in space. Even though Gordon coined the word "Stereophile," he wasn't one -- he's been waiting for multichannel music playback from before the birth of stereo.

Most of the other people I know who are enthusiastic about multichannel music formats are folks who don't have separate systems for music and home theater. Since they listen to their CDs with the same 5.1 speakers they use when listening to (and viewing) their movies, it strikes them as wasteful not to be hearing anything out of their "extra" speakers when listening to music.

It only makes sense, right? I do, in fact, see their point.

But here's my problem: the processors that assign the signal to the individual speakers were developed for HT applications. If you want to simulate jet flyovers (or bullets wizzing overhead) or pinpoint off-screen action, the existing processors do a great job.

But they do a terrible job at assigning music-enhancing signals to those "extra" channels. Part of this, of course, is that soundtracks are encoded to send discrete signals to each of the speakers, and the processor's steering is therefore simple, whereas there's no simple logarithm for assigning musical signal to the channels.

Hell, there's no consensus as to how to record multichannel music!

It seems so simple to me -- if you record the musicians with the front channels and the ambient sound of the space in which they are recording with the rear channels, you've pretty much dealt with the basics. Then all you need to do is balance the two signals (which is a lot harder than it sounds) and Bob's yer uncle!

However, almost no one has attempted to use multichannel this way.

Well, Peter McGrath has -- and his recordings have been revelatory every time I've heard them at hi-fi shows. And so has Gordon -- the high points of several of my visits to Gordon's house have been his playback of his and Steven Stone's recordings of the Boulder Philharmonic.

But I've attended something like ten years of multichannel demos at Sony, and every time they've proudly trotted out something like a Wynton Marsallis recording with Wynton's ensemble in front and a string orchestra coming out of the speakers behind me. I won't even sit with my back to the room at a restaurant; trust me, the last thing I'm going to turn my back on is an orchestra -- especially if it has trombone players!

A good illustration of this was Sony's demo at the recent High End Expo 2001. The room was crowded, so I ended up standing next to the left rear speaker. Sony played three pieces to illustrate multichannel SACD. The first was a Steve Epstein-produced Midori session. It was just about perfect -- I couldn't detect a thing coming from the rear speakers, but somehow an almost physical orchestra was playing Mozart in the room. But Sony proved this good taste was just a blip on the radar screen with their next two selections: one placed chorus members behind the audience, the other seemed to arbitrarily assign certain frequencies to the rear speakers.

Yeuccchhh! Pfui!

There are certain things up with which I will not put.

Okay, I hear you thinking, that's true enough for jazz and classical, but surely with material such as pop music, there are no hard and fast rules. Why not have effects coming out of the rear channels? Well, for one thing, most pop recordings simply recreate the experience of hearing the group or the singer in the real world -- so why spoil the illusion?

But I'll grudgingly concede that groups such as Radiohead and Massive Attack are creating new sonic worlds, and that all the old rules should consequently go out the window. In those cases, however, I'm sure that the artists will come up with something more clever -- and more meaningful -- than simply assigning certain frequencies behind the listener.

At least I hope so.