This sound model was inspired by the Prophet VS from Sequential Circuits.
The main design elements are four oscillators with selectable different waveforms, which are mixed together dynamically, controlled in time by a dedicated envelope, additionally processed through typical modules of subtractive synthesis (VCF, VCA with envelopes etc.). Each oscillator offers a choice from a set of different waveforms, resulting in rich timbres, changing over the course of time. Certain aspects of the sound can be modulated additionally by e.g. LFOs.
This implementation doesn't try to imitate the original instrument; it is only a showcase of the principles of the original design idea, presenting some examples of the possibilities. And it is a work in progress, a tentative draft.
The original instrument was a hybrid 8-voice synthesizer, combining digital sound generation with real analog filters and other elements. It used a total of four oscillators per voice with 127 waveforms (of which 32 were user definable) and dynamic waveform crossfading via a joystick and a special envelope generator, a manner of synthesis which was called "Vector Synthesis" by its inventors. One characteristic was the side effect of aliasing in some waveforms (see http://www.vector.synth.net/ or http://www.keyboardmuseum.org/ar/s/seq/pvs.html).
In order to complete the model among other things there are missing:
more modulation possibilities;
more sophisticated live modulation;
enhancement of the waveforms (better: replacing this poor example of only 13 waveforms by a more sophisticated design of an interesting set of waveforms).
The first sound in the soundfile represents the basic model; the second sound features an extended version (mainly enhanced by incremental parameters' modulation); and the third one shows one possible variant: replacing one multi-wave oscillator by a FM oscilllator (other variants could be: external input, samples, varying pulse oscillator ...).
Have fun and play.
Thanks to Kurt for some optimizations diminishing resource consumption.
-- EckardVossas - 25 Apr 2004
Meanwhile I worked over my first suggestion. In this revised version of the sound model the formerly used self-made "sample oscillators" are replaced by instances of the new oscillator prototype provided with Kyma X.10, making use of it's enhanced possibilities. Additionally the amplitude envelope (generator) is replaced by a more sophisticated one, going much further than the regular ADSR scheme. Moreover an broadly extended set of one-cycle waveforms is included too.