Procedural Audio and Binauralisation Using Max/MSP and the Unity3D Game Engine

Despite the reliance on procedurally generated (aka “generative”) audio techniques in early video game consoles, the nearly two-decades that have passed since the introduction of effective playback of recorded samples have seen few implementations of procedural audio in commercially released titles. The growing scale of modern games requires that we re-evaluate the potential for procedural audio to increase sound effect variation and decrease usage of memory resources. This paper describes the creation of procedural audio models for use within a modern, shooter-style video game setting. Max/MSP was used to create the models, which are triggered by a Unity3D-powered demo game environment, called “Bootcamp”, which has been altered to communicate with Max/MSP. In order to localise sounds, an audio binauralisation engine has been developed based largely on the principles of Lord Rayleigh’s Duplex Theory, in addition to simple front-to-back localisation. The result is a complete overhaul of the original demo game environment’s sound design, and a convincingly realistic one at that. My hope is that the features of this system can be integrated into a commercially released video game title, either by native integration of Pd into a game’s sound engine, or by hosting Pd within Wwise via a VST-style plugin.

How was MAX used?

All of the sound in the project and all of the panning/binauralisation is done in Max/MSP