In the early days of computer games the sound was created using simple synthesis techniques, but as the development of better processors and larger storage media, the development lead towards the use of wave-table synthesis, which has become the most used technique in current computer games. Since the introduction of the wave-table synthesis the development in audio creation and playback in games has stagnated.
One of the latest fields within sound synthesis is physically modelled, which holds great potential, within games and interactive environments, because of its more dynamic nature. An area in which very sparse research has been done is measuring the impact of physically modelled sound in computer game environments. This has lead to the following problem statement to be formulated: To which degree does physically modelled sound enhance physical immersion in first person computer games?
This project has analysed theories proposed by several authors within the fields: immersion, narrative and gameplay in computer games, audio in computer games. These fields and their different theories have formed an ontology for the project, upon which an application has been build. The application consists of a Half Life 2 modification, which makes use of the Nintendo Wii controllers, together with a modal sound synthesis.
To test the hypotheses derived from the problem statement, 24 test subjects were in a between subjects test, introduced to the aforementioned application, in two different scenarios: (1) one scenario testing 12 subjects, with the build in sounds from Half Life 2 and (2) another scenario with a modal synthesised sound of a crowbar, a weapon in the game, also testing 12 subjects.
The results of the test were processed by using a Student T-test with an alpha value of 0.05. In the question of how realistic the sound of the crowbar was perceived, there was no significant difference between the two scenarios. The test subjects definition of whether the crowbar sound was repetitive or dynamic showed a significant difference, in the terms of the physically modelled sound was more dynamic compared to the regular sound. In the questions of whether the subjects were able to judge the speed of the crowbar and if it was matching their hitting gestures, the T-test showed no significant difference.
It can be concluded that in the narrowest sense, wavetable synthesis can be replaced with modal synthesis, when creating a crowbar sound. In broader sense, what this project has shown is that in at least one case, physically modelled sound is as good as sampled sound. This has a lot of potential because it is more dynamic. In the broadest sense, physically modelled sound and sampled sound, when used in a first person computer games, yields the same level of physical immersion. There appears to be no technical or perceptual reason why physically modelled sound cannot, or should not, be used in computer games; therefore we suggest that further research should be conducted within this field.