Quote:
Originally Posted by kirkbross
My typical process is to create an Instrument track with Omnisphere (for example) and then create 8 MIDI tracks to send to the MIDI channels 1-8 of Omnisphere. i.e. 8 different sounds playing 8 different parts from a single instance of Omnisphere.
Is this the best practice in terms of computer resources or is it better to create 8 separate instrument tracks, each with an instance of Omnisphere, and have just a single sound in each one.
I ask because a friend said if they're all stacked in a single instance, they can only use one CPU thread.
|
In general and not just for Omnisphere but any multi-timbral vi it would spread the load better amongst cpu cores to have multiple instances of a vi rather than stacking all eight or whatever number of instruments into one vi. There are exceptions though.
Say I'm doing something orchestral where I have several instruments each sometimes playing both staccato and long notes. A trumpet line playing both staccato & longer notes would have one instance of a vi (Kontakt or EW Play in my case) with two trumpet instances with one MIDI line that play staccato notes going to the trumpet instance that has staccato samples. The other MIDI line that plays longer notes would go to that same instance of Kontakt or EW Play to the instrument that has longer samples. Now include a tuba with staccato and longer notes - that would go to another separate instance of Kontakt or EW Play. I would not stack every brass instrument in one instance of Kontakt or Play.