Interpolation error in waveform table lookup

Abstract: Waveform tables are an important tool for synthesizing sound, but they introduce error which results in noise. Error is affected by the spectrum of the signal stored in the table. Error is reduced by increasing the table size and/or by increasing the quality of interpolation. Both of these also affect the signal computation cost. Table sizes required for a given signal-to-noise ratio are computed for different interpolation methods and spectral rolloffs. Execution times are then evaluated. Non-interpolated oscillators perform the best, but only if the storage and computation costs of the tables are not an issue. This and other tradeoffs are discussed. 1