Additive Synthesis by Subtractive Resonant Filters

Juan Reyes (juanig@ccrma.stanford.edu)
Center for Computer Research in Music and Acoustics(CCRMA)
Department of Music, Stanford University
Stanford, California 94305

Abstract:

Resonant filters can be fine tuned to a very narrow frequency band thereby isolating a tone even from a non-pitched sound source. ``Maxf.ins'' is Max Mathews (mvm) new filter (2002) described as a High-Q, 2-Integrator, filter with two poles, and one zero at the origin. This CLM implementation renders equal tempered frequencies integer & just scales out of a wide-band input signal. The filter might be used for Modal Synthesis but also might be Additive Synthesis in which a resonator is initialized to generate the exponentially decaying sinusoids at the desired phase. Different states which are bound in parallel are defined in a structure which contains different frequencies and tunings for resonant modes. In this algorithm the filter is recurrent over the source signal by iterating the number of desired frequencies in a state. States can be defined as containing at least one frequency up to the CPU processing power. We have tested as many as sixteen states.