Computer processing of audio signals by exclusion filters
Robert C. Maher · The Journal of the Acoustical Society of America · 1990
A common task in computer music and electroacoustic signal processing is additive mixing of two audio signals. If the two input signals contain discrete spectral components, their sum will typically contain amplitude beating and other interactions between pairs of components with similar frequencies. A new method is described that suppresses spectral interactions during mixing by deriving a time variant “exclusion filter” from the short-time spectrum of one of the signals in order to prefilter the other signal. This technique allows one of the signals (dominant) to pass through the mixing process with little modification, while the other signal (secondary) is prevented from interaction by attenuation of its conflicting spectral components. The exclusion filter is specified in a flexible manner, which can include such psychoacoustic criteria as critical bands. Audio examples will be presented.