Vector average and divisi

This is not as critical to me as the cue request but its a nice idea for the coming age of orchestral vsti as well as synths and midi devices in general.

For Synful, WIVI, DVZ, Garritan, Samplemodeling etc the trend is to create sections from individual instruments. I feel this is the way forward.

The variations on human performance in section playing could be handled in Synfire, more consistently across vsti's and with compositionally informed sophistication that is not likely in a VSTI.

This would also put the artistic decision of how tight a section should be, the variance in dynamics or how in tune, back into the compositional software, not the sound generator, to benefit from the Synfire model.

One idea would be to record two passes of a Figure, with all of the cc controls performed for each, and then select how many voices should be sounded varying the parameters based on an analysis of the difference across all parameters for both performances. From the high and low limits of two performances 24 violin performances could be derived.

An analysis of the rhythm of peaks and shape could be handled similarly to figures (Think Melodynes display with midi notes atop pitch graphs, which distort based on the notes offset.) The interrelation of cc and figure vectors could come in to play. Also cc values could be understood based on their nature, for example; pitch and expression would be analyzed for the rhythm and shape of peaks and troughs, where as attack noise would be understood to have no rhythmic meaning on a single note scale.

Perhaps these analysis could be saved and then applied to single performances in other parts.