LucasArts Mentioned in Contemporary Music Review

[MarioColbert]

So here I was reading Contemporary Music Review from February of this year. The issue is titled "Generative Music" and deals with algorithmic and otherwise "coded" computer-music. While I understand that computer music research field renders me the laughing stock of these boards and likely to get the rest of you lot to point at me with your index fingers and laugh and laugh, the reason why I'm embarrassing myself with such private information is because LucasArts was mentioned in the article! TWICE!

 

First off, they mention iMUSE: (read only the part in BOLD if you like quotes out of context)

 

A related form of transformational algorithms used in game composition has been the smaller parameter-based changes that have been programmed into larger sequences of music, defining start and end points, jumps, repeats/loops, and so on. In these cases, a composer creates musical sequences, and then conditionalizes those sequences, making decisions involving which segments should play continuously, change, be enabled or disabled, loop, converge, swap instruments, branch, jump, and so on. Markers, or decision points, indicate places where changes or branches in the performance may occur based on a condition in the game. For example, the patent for iMuse, a music engine designed at LucasArts in the early 1990s, describes the process as follows:

 

… fight music, rather than playing along unresponsively, can be made to change [the] mood of the game in response to the specific events of the fight. For example, certain instrument parts can signify doing well (for example, a trumpet fanfare when a punch has been successfully landed), while others can signify doing poorly (for example, staccato strings when the fighter has taken a punch) …Also, it may be desirable to transpose … the music as the fight reaches its climax. This can also be done either immediately under entertainment system control, or by hook message (if it is necessary that the transposition occur only at an appropriate point in the music sequence). The resulting fight music will change mood and character along with the intensity and excitement of the fight, but in a smooth and aesthetically natural way, much like the music would follow the action in a feature length motion picture. (Land & McConnell, 1994, p. 5)

 

Although the game engine cannot necessarily always predict what actions a player may take, the illusion can be created if music can respond to various parameters. In the above case, for instance, the music could be altered according to player and enemy's health meters, such as a simple 'if player's health is more than enemy health, then play winning music. if player's health is less than enemy health, then play losing music.' The music can be controlled by a jump cue—'jump to winning section of song'—or could be altered on the basis of instrumentation changes—'add trumpet fanfare'.

 

I followed up the Bibliography bit of this citation and found a full-text public domain version of the original cited document which is LucasArts patent for the iMuse system - http://www.patentstorm.us/patents/5315057/fulltext.html (pretty fancy stuff, if you're into this sort of thing).

 

But the article goes on to mention a game that I've forgotten about for ages until it was mentioned in a recent LucasArts roundtable from 2003 (with Gilbert et al.) or was it in the Tim Schafer's recent post about his LucasArts job? Anyway, here's the juicy quote: (again, the bold emphasis is mine)

 

In addition to these transformational algorithmic-controlled examples of game music, there have been several projects to incorporate what Wooller et al. (2005) define as algorithmic generative music. Lucasfilm Games' Ballblazer (1984), for instance, employed what composer Peter Langston termed the 'riffology' algorithm, in which dynamically weighted choices for parameters are decided by the computer:

 

such as which riff from a repertoire of 32 eight-note melody fragments to play next, how fast to play it, how loud to play it, when to omit or elide notes, when to insert a rhythmic break, and other such choices. To choose the next riff to play, the program selects a few possibilities randomly (the ones that 'come to mind' in the model). From these it selects the riff that is 'easiest' to play, i.e. the riff whose starting note is closest to one scale step away from the previous riff's ending note. To decide whether to skip a note in a riff (by replacing it with a rest or lengthening the previous note's duration) a dynamic probability is generated. That probability starts at a low value, rises to a peak near the middle of the solo, and drops back to a low value at the end. The effect is that solos start with a blur of notes, get a little lazy toward the middle and then pick up energy again for the ending. The solo is accompanied by a bass line, rhythm pattern, and chords which vary less randomly but with similar choices. The result is an infinite, non-repeating improvisation over a non-repeating, but soon familiar, accompaniment. (Langston, 1986)

 

Langston claims that while the song passes the 'is it music' test, it fails to pass the 'is it interesting music' test, 'because the rhythmic structure and the large scale melodic structure are boring. It appears that for music to remain interesting it must have appropriate structure on many levels' (Langston, 1986).

 

Again, a quick look at the bibliography points to the source: http://www.langston.com/Papers/2332.pdf (WARNING: this is pretty effin' technical) but you can search the document for LEVINE84 (one string) to find a rather technical (but amusing) mention of LucasArts' forward-thinking musical strategies for Ball Blazer.

 

I'm excited about this find for several reasons. You see, in my professional world, the games I grew up with are hardly ever a topic (although in case of Schafer's latest offering, a few fellow doctorate students have mentioned interest and one even purchased it himself). Not only is this "relevant to my interests," but apparently LucasArts was all kinds of proper hardcore about technological advances necessary to create interesting sound occur without occupying more than a megabyte of disk space.

 

So even if you don't give a rat's ass about all this academic music nonsense (who can blame you apart from ME?) I thought you'd be excited to know this stuff about the allmighty LucasArts of the days past.

 

 

P.S. I'll gladly post the CMR article as it's a good read even for non-music people, but first I have to remove this annoying title page that states where I go to school. If any of the forum admins think that posting a link to an academic article is a bad idea due to some copyright violation possibility that I'm not aware off, please let me know and the crisis will be averted.

[s-island]

http://www.uspto.gov/ - if you go here and search for lucasarts, you can get the whole patent document, including the images.