To produce individual tracks, as apposed to "Entire Mix"

Hi.  Hope everybody is doing great music.  I got a question:
When bouncing to tracks (Producing the audio track out MIDI channels) I need to produce individual audio tracks as opposed to the "Entire Mix" so I can manipulate them separately.  This is no problem if I load multiple instances of the instruments and use a single instance for every MIDI track.
 
The question:  Is there any way to produce individual Audio Tracks without loading multiple instances of the instruments that otherwise are multi-timbral in Sonar Platinum?
 
Thanks a lot.

It depends in which VSTi you are using, as some are limited by the VSTi itself (for example, TTS-1 will take 16 MIDI channels in, but only make 4 stereo outputs). The easiest method is when inserting the synth to check "all outputs stereo" (or mono) when inserting, but after the fact would need to add audio tracks and reroute the synth (or insert a new synth and move the MIDI data).

Scook posted a very nice screenshot of the output options for the TTS-1 in this post.
 
The TTS-1 has 4 outputs (you can think of like busses), so the max you could get if you truly want individual instruments is 4 per instance. If you enable "All Outputs: Stereo" when inserting the TTS-1, you will see all 4 of these right away without further need for routing.
 
Having said that, it may also be overkill to do so; as it is also common to have similar instruments grouped (hence the "buss" concept, so in reality they may get very similar audio processing depending on the situation). 
 
Another option, is the TTS-1 is awesome for doing composition, but may not be the best "end result" you want. I will often compose with the TTS-1, but then insert another soft synth and Shift-drag/drop the MIDI data from the TTS-1 to the other VSTi for further editing.
 

Basically, some soft synths have incredible sounds, but can also come with a price - CPU usage. People often find this out the hard way when they create a project template with every synth they own pre-loaded. This is often the result of internal effects to the synth.
 
When working with MIDI data, it is simply the notes of a composition, and TTS-1 will play them for the purpose of hearing how things sit in the mix. But you may want the sound another synth produces.
 
From strictly a CPU perspective, it can be easier to compose in TTS-1, then insert a synth to "replace" a given channel. At that point, you can mute the TTS-1 channel and copy the MIDI data into an inserted synth, so is now being played by the other VSTi. If things get too CPU intensive, the new synth can be "frozen" which essentially makes it a simple audio file (which is played back easily). One caveat here is that not all synths map note-for-note, so some tweaking/transposing of that data may be needed using this approach (for example, the SI bass is different by an octave IIRC, so needs to be transposed to sound the same as the TTS-1).
 
The same concept applies to putting a lot of audio FX on a track... each time the song plays, the CPU has to process each one. Minimizing FX used, and committing them (via bouncing) will lower CPU usage. Also note that a bounce (render) creates a new audio file, but does not delete the old one, but may make it hard to find if the project is not saved with a new name (see below).
 
The variations to this are pretty endless, so really depends on your preference. But one final point is whenever bouncing/rendering, I often make a point to save the project with a new descriptive name (in the same folder), such as "[project x] guitars bounced" so that I can drill back into the project history if needed at some point later on. This really is for audio file concerns though.
 
Edit: The above technique is also handy if you get a MIDI file from online. TTS-1 will load it, but you may need other synths to make it sound like the original.