Farnell (2007) describes procedural audio as,
“non-linear, often synthetic sound, created in real time according to a set of programmatic rules and live input.”
This means sound is created generally synthetically whenever the player of the game initiates a set of conditions within the game world, for example the player swings a sword. This could be done with multiple recorded assets, but this approach uses valuable memory and has to be recorded leading to additional cost.
With the constraint of memory ever present in games, how can we further relieve these constraints on audio variation? Procedural audio is one such way. Fournel (2010) tells us that procedural audio is used:
- due to memory constraints or other technology limitations
- when there is too much content to create
- when we need variations of the same asset
- when the asset changes depending on the game context
Procedural audio for sound effects is real time sound synthesis, one example of this is SoundSeed by Wwise.
Wwise claim that by using Soundseed developers never have to compromise on variety. The Soundseed Impact modeller and plugin demonstrated in the above video can take just one recorded source file, analyse its content and then create variations of the sound at run time. It is however possible to create sound from the ground up in Soundseed as well as in other similarly purposed products, this gives the ability of creating original unheard sounds if desirable, this is ultimately the more time consuming and technically difficult.
As well as impact and resonant sounds, Soundseed can also create wind sound which would have a large memory footprint. Fournel (2010) explains that good candidates for PA include:
- Repetitive (e.g. footstep, impacts)
- Large memory footprint (e.g. wind, ocean waves)
- Require a lot of control (e.g. car engine, creature vocalizations)
- Highly dependent on the game physics (e.g. rolling ball, sounds driven by motion controller)
- Just too many of them to be designed (vast universe, user-defined content…)
If we look again at Paul’s Pure Data patch from last weeks blog he demonstrates wave modelling here:
he demonstrates building sound from the ground up. And because the sound is synthesised it can be made to sound different every time it is played as he demonstrates for the game Sim Cell.
At present PA does not provide an all in one solution, Fournel (2010) explains that it is excellent at recreating sounds like wind, water and impacts but for other sounds there is a lack of sound designers with the technical knowledge for implementation. Hopefully in the coming years development teams are better trained in this area as Fournel (2010) suggests.
Farnell (2007) An Introduction to Procedural Audio and its Application in Computer Games. [Online] Available from: http://obiwannabe.co.uk/html/papers/proc-audio/proc-audio.html [Accessed 25 October 2015].
Fournel (2010) Procedural Audio for Video Games, Are we There Yet? GDC Vault [Online] Available from: http://gdcvault.com/play/1012645/Procedural-Audio-for-Video-Games [Accessed 25 October 2015].
LostChocolateLab (2010) Audio Implementation Greats No’8: Procedural Audio Now. [Online] Available from: http://designingsound.org/2010/09/audio-implementation-greats-8-procedural-audio-now/ [Accessed 25 October 2015].