This is the first attempt in creating a dynamic spread generator engine, in which you can create independent spreads, each one driven by a custom generator (an LFO, a simplex and a random+queue).
Once created, you can set each spread bin size.
To achieve the modular generation structure I used delegates, trying to understanding their pros and cons.
What do you think of using delegates in this way?
it seems they don’t allow stateful patching? is it right? this is a limitation in designing any sort of spread generator…
do you suggest any other approach to the task?
a desired thing would be to keep separated the design of the engines (LFO, simplex, any sort of generator) from the dynamic spread dispatcher part, in order to be able to add new engines (adding new delegate regions to the cons in the EngineDelegates patch) without touching the rest.
the patch looks totally valid. switching the delegate to generate another set of values is the right method for this purpose. it’s a bit like exchanging a node in the patch during runtime…
a delegate is basically the same as a util operation with the only difference that you can send it over links. so there is no state like properties or process nodes that can be remembered and re-used in the next call.
in one of the upcoming vl releases we will support interfaces, this will be another method that can solve this and similar problems.
Tnx tonfilm,
I’ll stay tuned for interfaces implementation then. If they work similar to dx11 class implementation they could be really the best tool for this scenario…
