I have a habit from when I did not have such a hefty computer.  If I wanted to use more than 1 sound from a soft synth - I would say - do 1 pass.. clean it up.. then freeze it.  I would then clone the audio and associated midi.  I would then unfreeze the synth - delete the midi - change the sound in the synth - then move forward.  I don't mind this process - and it would help preserve the flow - if say - I am using a long pad that evolves over 30 bars etc- where the synth would not recognize a note if I would start after the note were triggered.  However - I know that people with more soft synth experience than I here on this forum may have a much more sophisticated technique, and I also want to make sure I'm reaching maximum fidelity in my soft synth work... as I do not know if the freeze command uses maximum fidelity - or not
 
Thanks
Dave
 
 

The workflow is a bit strange. You do not need to delete anything generally: you can keep gazillions of tracks in the project with no appreciable load on your computer if they are frozen, and just hide them if they are in your way. My understanding is that the freeze is rendered at the full fidelity set in the program, just as if the track is playing live. When you render the track to a wave file on export, you generally lose some of the information that the program uses internally for processing. 
 
edit:
 The bit depth of the freeze is under user control: Edit > Preferences > File - Audio Data, in the Render Bit Depth field.

I would suggest that you acquaint yourself with Sonar's synth rack. Just load your VSTi's in there and point your midi tracks to the appropriate synth/channel. Easy peasy and you should be able to run a fair number of synths/samplers unless your computer has insufficient resources.
 
As for retriggering a note where the note on event is prior to your start point, there is a feature request to add that, but Sonar still ignores notes like this unless you start playback on or before the note on event.
 
Regards,
Dan
 

If I have a project that I want to use 2 different synth sounds from, and the synth is not multi-timbral (can only play 1 sound at a time, whereas multi-timbral is where a synth can play multiple sounds on multiple midi channels at the same time), then I simply insert an additional instance of the synth into the synth rack.
 
So, each instance of FM8, for example, would have a different sound loaded, and would have its own audio track(s).
 
With regard to the midi, if both sounds are to play the same midi data, clone the midi track and point that midi track to the 2nd instance of FM8, and have the audio track(s) for that instance point to the output audio from that 2nd instance, and then both sets of midi tracks will trigger their own FM8 sound, with one sound coming from each instance of FM8 in the synth rack.
 
I hope that the above is understandable - I AM on my very first tiny fledgling coffee, so I don't even know if I am typing in English yet.  :)
 
Bob Bone

u can also sample VSTi pugins,  DirectWave https://www.youtube.com/watch?v=-ROoPtv9g5Q knows it automagically ofc. the inbuilt VSTi fx units will be lost this way but it no needs freezing at all