Getting to grips with Endpoints

I’m trying to use the new endpoints feature with ARIA Player and Garritan GPO5 library.

Endpoints are great for setting up ‘standard’ groupings, like SATB+org, but as soon as the scoring varies from the configs that I’ve set up, I get another VST Instrument with a complete duplicate set of samples.
So, for instance, if I create a VST with choir, strings, 1 oboe, trumpet and organ, and then I have 2 oboes in my score, I get another complete ‘orchestra’.

I can mitigate this somewhat by creating lots of different endpoints within the playback template: SATB, SATB+org, SATB+org+strings, SATB+org+strings+woodwind; SSAATTBB.

What’s the best way to manage scoring variation? Does it matter (in terms of performance, memory, power, etc) if I have 16 VST Instruments each with one sample, rather than 1 VST with 16 samples?

From a performance perspective, the resource usage will differ depending on the way the VST instrument itself is implemented. Without asking the manufacturer of the plug-in their recommendations, your only way to determine this would be by experimentation.

I think probably a decent way to approach this would be to strike a balance, e.g. create one endpoint configuration for double woodwind consisting of one instance of your multi-channel plug-in, one for brass, one for strings, one for voices, etc. You can then create different playback configurations that combine these configurations together. By their nature, because they are presets, you’ll always run the risk of loading sounds you don’t need, but that’s the price you pay for presets.