Note: This is a piece I posted in 2010, lamenting what I percieved as a tendency to neglect truly integrating a soundtrack with a game in the way that some games very ambitiously attempted in the early 1990s. On reflection, I think that this is a little bit of an over-simplistic way of looking at it, and games are actually doing quite a lot of interesting things with music. But I do think there are still a few interesting points here, and I include it for those, and for the video demonstrations of old iMUSE technology, so I've included it, with my own comments, in italics.
If someone were to ask me what I think the most technically sophisticated game was in its visuals, I would probably choose something quite new. If they asked me for some of the most refined examples of game design, recent examples spring to mind just as readily as older ones. It’s puzzling to me, then, that when I consider video game music, I have to cast my mind all the way back to 1991 and LucasArts’ iMUSE system before I get something that approaches what I would consider state-of-the-art. Could it be that I’m giving too much credit to these old games, too little credit to newer titles – or am I onto something? (I think perhaps the former. Monkey Island 2 is state-of-the-art in some ways, newer games in others.)
In 1991, LucasArts released Monkey Island 2. Not only was it a popular and successful game in its own right, but it was also the first to showcase a system called iMUSE and arguably the most complex implementation of it. But why this makes theMonkey Island 2 soundtrack so advanced takes some explaining. Imagine how this scenario would be scored in a film: A bad guy walks in, confronts our hero, then leaves. The scene has a set length, so the composer writes some bad guy music appropriate for the length of the scene, then has the music end appropriately as the bad guy leaves.
Next, consider how this works in a standard video game. The bad guy enters, and the game cues up the bad guy music. Because the scene is interactive it could be different lengths, so when the bad guy leaves generally the music either stops or is gently cross-faded with whatever the background music happens to be. At best, a short ‘bad guy leaves’ sting will play, interrupting the bad guy music before the ambient music resumes.
iMUSE was different. Concealed in the bad guy music are special codes, every couple of seconds. They are placed at every bar or so (the natural rhythmic transition points in music). When the cue for the bad guy to leave happens, the bad guy music continues to play until it reaches one of these codes. At this point, it jumps to one of several endings that have been written, dependent on what was going on in the music when the bad guy left. The result is that the music always ends properly, no matter what you do, and there are no out-of-sync interruptions or awkward cross-fades involved (see the first video, below).
Not only can iMUSE seamlessly end music but it also effortlessly transitions between different pieces of music as you move from area to area, using the same technique except with little musical bridges to connect each piece rather than just endings (see the second video). Other tricks it could pull include bringing instruments in and out of the mix depending on where the main character was standing or what he was doing. All of this adds up to a musical score which does not let the game cut up, yank and pull at it as needed, but coexists so that the music flows around the gameplay in an utterly convincing manner. (As long as the composers are willing to wrote enough transitional bits for the soundtrack to keep up)
Michael Land, Peter McConnell and Clint Bajakian’s score for Monkey Island 2 is huge, ambitious, full of terrific tunes and one of the unsung masterpieces of gaming but all of that is almost beside the point. The real question is why have so few games since attempted to marry gameplay and music to such a level of sophistication? Artists and engine programmers are constantly finding new ways to make textures and animations more believable and games that treat these sloppily are rightly criticized. Yet nobody seems to want to really get to grips with a truly dynamic, interactive score to anything approaching the extent that those three guys managed almost twenty years ago.
One of the problems is that back in the heyday of iMUSE, most game soundtracks used MIDI. A MIDI file is nothing but a series of instructions that are interpreted by the sound card. What sounds actually play depend on your particular sound hardware and software. In the videos, for example, I was using quick ‘n’ dirty Windows MIDI which does the job but, as you can probably hear, doesn’t exactly have gorgeous instrument sounds. Because the MIDI is simply a set of instructions, though, it’s very easy to tweak it in a way that sounds seamless. It would be far trickier to record the same sort of seamless transitions and endings in pre-recorded, digital audio. In order to pull off the trick of bringing instruments in and out of the mix (below) the sound engine would have to have some sort of built in mixer with each instrument playing in separate channels.
These sorts of hurdles are by no means insurmountable but they do take time and money to implement and such techniques would be a little more resource-demanding on whatever system the game is running on. I suspect that most developers have long concluded that the current method of using cross-fades or simply giving up and having jarring musical transitions is “good enough.” The idea that there is room for improvement never even crosses their minds. Isn’t it time we started to question that, though? (On the other hand, they might be right, that in most cases this is good enough. But I would maintain that iMUSE was a very worthy experiment and something like it for modern soundtracks is well worth exploring)
Production values in games have never been higher and this is a trend that will no doubt continue. As far as visuals go, nothing is ever “good enough.” There’s always another improvement to be made and the technology is constantly being developed to meet the challenge. It’s time for game composers to adopt a similar mentality, to take more care over how their music is implemented in the game and demand the right tools to make their ambitions achievable. (Of course, it's rare that they're actually in a position to make these sorts of choices and demands, but I like my idealism) In the meantime, it’ll be interesting to see how well LucasArts have managed to implement the Monkey Island 2 soundtrack in the upcoming Special Edition. (It turned out, pretty well!)