Stephan SchÃ¼tze looks at what modern generative audio methodologies offer games designers
Generative audio remains as powerful an option now as it was ten years ago when I first applied it to a PlayStation 2 game.
Its many benefits include reduction or removal of repetitive sound effects, more efficient use of memory, and if done properly, a better sounding audio environment. Generative audio continues to evolve as a process with the availability of new tools.
One of the most exciting upcoming tools is FMOD Studio, and I have been fortunate enough to get access to an early copy of the software to discover a little of how this technique is developing.
Generative audio is the process of using small audio building blocks to create sound effects in real time. Essentially you take the smaller elements used to create a sound such as an explosion, and enter those elements directly into memory.
When the explosion sound is triggered, it is assembled in real time. Usually, this process utilises a series of sound-files that offer variable components, increasing the variety of the generated sound.
The beauty of this method is that once a sound file is in memory it can be reused infinitely. The sound of a shotgun is used for its obvious purpose, but then pitched down to simulate a grenade, or pitched and combined with other sounds to create an explosion or collision impact sound.
In FMOD Designer, for birds in a tree I would utilise a series of short bird calls and trigger them with defined delays, pitch randomisation and suitable polyphony. This was effective, but still required a reasonable number of short birdcall sound files, adding to memory. FMOD Studio’s new functionality allows for even smaller initial components. Many creature vocalisations are the product of a series of similar tonal sound elements; it is possible to generate a reasonable likeness of a bird or animal call from the smallest component of the vocalisation.
A group of insects in an environment is much like a group of people, with a constant level of conversation between those in the area. With the current generation of tools we can isolate a single insect ‘sentence’ and use the generative audio functionality to simulate a group of insects in a region. We could possibly even break it down to a single ‘phrase’ of an insect to create the effect of a larger group.
Two elements within FMOD Studio take things further. An Event can contain both a Timeline Track and/or a series of Parameter Tracks. This means Events can have a more complex interaction of associated sound files. The second element is that Events can be nested within other Events and controlled both up and down the nesting hierarchy. They can also be controlled by Events outside of the nesting chain. The possibilities these functions provide allows you to not just utilise an insect call down to the ‘word’ level, but further still to a ‘syllable’ level.
FMOD Studio is still in development, but it appears generative audio is about to take a large evolutionary leap.
GENERATIVE AUDIO ON THE MOVE
I am reasonably confident that the full version of FMOD Studio will allow me to create the environmental audio of a massive game world such as a Skyrim or a Borderlands, using the smallest sound elements and the generative audio process to build a dynamic and repetition free audio environment small enough to work on a mobile device.
This will be a tremendous benefit for the rapid growth of development of triple-A projects on mobile platforms, as well as an incredible improvement of efficiency for larger console audio, supporting the creation of more engaging audio environments.
For me, video games audio has already surpassed many other media formats in its production complexity, as well as in its potential for excellence.
Generative audio is an element of game production that will become more and more the norm as we strive to create striking audio environments for our players.
The promise that new tools such as FMOD Studio are offering has me excited, as I honestly have no idea what I will be capable of creating in the next 12 months.
The evolution of game sound and generative audio will be directly influenced by how far we want to take it.