As a kid, I was not allowed to gorge myself on video games to the extent of many of my peers; my mother simply would not allow it. I did gorge myself on television, but still less than some. Even so, when today I happen by some bizarre turn of events to catch a glimpse of a cartoon or video game, I'm typically most surprised at (a) the turn towards realism, and (b) how profoundly aesthetically unsatisfying this is to me. It's an old man's gripe to be sure (I'm not yet 30, but in the technological world, that's middle age), and I've heard enough of them directly from old men to be wary of committing the same fallacies. My gut reaction is nonetheless remarkably consistent. Modernist though I claim to be, perhaps I'm finding for the first time some appreciation for the advice of so many conservative composition teachers that imposing limitations on one's process can be beneficial to the outcome. There are more than a few vinyl hoarders and NES players (and conservative composers) who would agree, no doubt cherrypicking their evidence with the utmost caution and backtracking appropriately when confronted about their iPhones.
The reason I bring this up here is that I'm coming to view notation software playback as one of these areas. When I upgraded to Sibelius 6 in 2009, I had been using version 2 since it first came out (yes, that's kind of a really long time). There was quite a bit to learn, a lot of useful new features, and a few real pissers. (The chord symbols! Barf...) The biggest challenge to this day, though, has been the built-in sounds. They are much more "realistic" sounds than the old general MIDI sounds I had become very accustomed to, by which of course I mean that it would now be much easier, possibly even a foregone conclusion, to identify by ear the instrument they purport to represent. (Forget the specific instrument; with the old sounds, you sometimes wondered which instrumental family was in play.) For whatever reason, though, I find them much more difficult to work with: the timbral whole is still less than the sum of its parts.
In some cases, notably the tuba, this is because they've essentially built mistakes into the samples:
Yes, there's a better than average chance that your garden variety community band tuba player will wobble slightly on a low A before the pitch stabilizes, but seriously guys, let's just shoot for the stars next time and pretend that tubists are at least theoretically capable of emitting a steady tone for more than 2 beats at a time.In other cases, I undoubtedly struggle because I spent an incredibly long time working with the old general MIDI sounds and hence got very accustomed to interpreting them. Space, balance and blend have always been the achilles heels of notation software mockups of through-composed, acoustic music, and I don't think the present results are any more accurate despite representing an obvious attempt to improve in just these areas. I'm left to wonder if I am, in fact, just getting old, or if we had not actually stumbled on a semi-optimal degree of reality, unbelievable as that would have been at the time, in comparison to which the next rung of progress actually looks regressive. Perhaps a virtual reality that is obviously fake would be more useful here than one with loftier aspirations and spectacular failures.