Generative music in games

„Cheap sound in many variations!” This may be expected from computer-gereated music in the best case. A computer doing a musician’s work could save money and discussions and hours of work. An artificial intelligence takes control of human decisions.

This may sound too much of a futuristic wonderland story. In early 21st century computer do create music but most people tend to love human-made music and would never long for artificial sounds. Just to think of the legions of Rock lovers out there, and Jazz music lovers and Singer- Songwriter lovers and so on. N ot everyone shakes his or her body for hours along electronic beats.

It may be an art to let an algorithm sound human to serve a broader audience.

Let’s shed some light on the musical reasons for taking over a computer.

First of all, a composer or game music designer has to make dozens of decisions. All musical elements are submitted to decisions: melodies, tone-pool, rhythms, tempo, pace, hits, swells, breaks, sounds, mood, emotion and so on. What if these decisions should be controlled by the player’s progress and situation? What may sound queer for a shooter may be a fabulous idea for a puzzle game or an evolving strategy game.

The world’s oldest musical instrument, not completely driven by human will are the windchimes. The tone-pool is set (by the tube’s lengths) but the rhythm and variatiopn is completely out of our hands – controlled by thermodynamics. Still, the instrument is predictable, it would not play different tones out of the blue. Solely the compository decision of rhthm has been given out of hand.

Probability-base systems can alter compository decisions and reach unknown musical territories.

We dream of the fact that a game reaches the “plus-zone”, the complex region which cannot be designed, the zone of ultimate personality, when the complexity of the gaming algorithm and the real player’s situation (his or her psychic disposition) interact and fuse into a distinct experience which cannot be predicted. What if the music had the same potential in a game? Not being fixed to set “songs” the sound could possibly reach uncharted territory and peronalize the gaming experience.

British composer Brian Eno has dealt with chance-driven music since many years. One of his works is the soundtrack of “Spore”, where he fed a system designed by programmers with sounds and patterns.

Lastly it makes a huge difference how sensible and creative an algorithmic system has been designed. Many hybrid forms of probability based systems and traditional approaches have been developed, also in techno and club music. One of them is seen here, in an example of a popular music software.