Computation and Memory Tradeoffs with Multiple Wavetable Interpolation
Andrew Horner · Rare & Special e-Zone (The Hong Kong University of Science and Technology) · 1996
A multiple wavetable interpolation technique for matching acoustic instrument tones is introduced. The method uses a genetic algorithm to optimize the parameters, where the user defines the number of oscillators and wavetables. Previous work on wavetable interpolation has only worked with a pair of oscillators, whereas previous wavetable matching has only considered continuously active wavetables (that is, equal numbers of wavetables and oscilla-tors). Results are presented using this new method to match a trumpet, a piano, and a wahwah muted trombone. It is shown how the different combinations of wavetables and oscillators perform, especially in comparison with simple wavetable interpolation and wavetable matching. It is also shown how to use these comparisons to help determine the best combination of oscillators and wavetables for systems with various memory and computation constraints.