Jari Kleimola
Design and Implementation of a Software Sound Synthesizer
Master's Thesis
Abstract
Increased processing power of personal computers has enabled their use as real-time virtual musical instruments. In this thesis, such a software sound synthesizer is designed and implemented, with the main objective being in the development of a composite synthesis architecture comprising several elementary synthesis techniques.
First, a survey of sound synthesis, effects processing and modulation techniques was conducted, followed by an investigation to some existing implementations in hardware and software platforms. Next, a formal object-oriented design methodology was applied to capture the requirements of the implementation, and an architectural design phase was carried out to ensure that the requirements were fulfilled. Thereafter, the actual implementation work was divided between the reusable application framework library and the extended implementation packages. Finally, evaluation of the results was made in form of sound and source code analysis.
As a conclusion, the composite synthesis architecture was found to be relatively intuitive and realizable. The generic object-oriented design methodology applied appeared to be well suited to the design of sound synthesis systems in general, but was considered to be too laborious to follow in every detail. The implementation work benefitted from the properly done design phase, however. The relative amount of man machine interface code compared to other subsystems was still surprisingly large. The timbral dimension of the realizable sound palette appeared to be quite wide, and the quality of the audio output was comparable, or even better than that of the existing implementations.
Keywords: audio effects, musical acoustics, object-oriented design methods, software framework, sound synthesis