Digital Audio Applications of Massively Parallel Architectures
In this paper, we will discuss novel uses of parallel processing in the analysis and synthesis of digital sounds. The paper begins with a brief discussion of parallel processing terminology and concepts, and then focuses on two specific areas of application. The first application is digital audio data compression. Two methods will be discussed, one involving a simple pattern-matching algorithm, and one based on frequency-domain analysis using the FFT. The other application is real-time sound production using complex mathematical functions. This technique is used to create a sophisticated extension of the traditional FM synthesis algorithm. The second half of the paper details the implementation of these algorithms on parallel architectures. Software simulations of the algorithms of the algorithms were developed to produce a variety of recorded examples. These experiments demonstrate the feasibility of real-time implementations in cost-effective hardware, such as might be used in next-generation synthesizers and digital audio equipment. The paper concludes with some suggestions of additional application areas for parallel processing in the audio field.
This paper costs $20 for non-members, $5 for AES members and is free for E-Library subscribers.