Programs vs Combi/Multi


I’m just starting learning JUCE and aim to create a synthesizer plug-in to re-create a classic rompler.

For this project I’d like to implement the using of layers/split in order to emulate multitimbral configurations (e.g. ‘Combinations’ on KORG workstations, or ‘Performances’ on Yamaha machines).

Maybe my question is stupid (I’m a newbie) but I can’t figure out what’s the proper way to implement this.

I see two possibilities:

a) Use a single Synthesizer object and a huge array of voices, then just assign voices to different patches/layers.

b) Use one Synthesizer object for each layer and keep things separate for each program.

Any thoughts on this?

Thank you in advance