ideas for interactive midi phrase synths + arps

i was just reading about the process of the older cubase IPS phrase synthesiser, and it seems it's a good way forwards in reaktor away from the basic 12 step sequeces etc... lfo variation generators, step distances, scalers etc etc...

good area to research for those interested in algorithmic music is algorithmic arpeggiation.

reaktor has alot of really advanced modules for this kind of thing, like unison spread and set random (to set random seeds (?))
anyone know what these can do?

if you made a complacated phrase synthesiser, would it be possible to configure it to play several different imported instruments all in the same ensemble, or is the in ensemble midi restricted to having only 1-1 2-2 style midi track configuration...

 

I have a sequencer I've been toying with for years. The idea is to let you work at a higher level of abstraction, or down in the details. Play in a chord progression (timing unimportant) and it can generate melodies basslines and chords, always within the progression. Play in a rhythm, draw a general melodic shape. Each piece can be manipulated independently. There's a "strummer" meant for working with my modeled guitar, and a "raindrop" generator for semi-random arpeggios/grain clouds.

One reason it's languishing is I've been a bit frustrated by the crudeness of Reaktor's timing measurement and reproduction modules. Is there anything more accurate, or have any improvements been made in recent Reaktor versions? For example, it appears that the Timer, Hold, and Event Delay modules all have a resolution equal to the audio latency, much coarser than, say, the control rate. Also, they are disrupted by various kinds of mouse movements.

By the way I played with the unison spread module the other day and it didn't appear to work: all voices came out with the same value, regardless of inputs and settings.