slartabartfast
The disparity in replies seems to be due to how you interpret layering. You can record/play MIDI from a single controller simultaneously onto separate tracks, each feeding the input of separate softsynths, and monitor the output to two separate audio tracks simultaneously. That is a lot more flexible than merging the sound into a single track, which is what some people seem to think layering is. That merged output is what you will get from a keyboard synth that merges the sounds into a single audio output, but then you lose the ability to tweak or assign automation to the two voices/synths independently. I do not see why anyone would prefer merged tracks in a system that permits unlimited tracks to be recorded.
In my case I would define layering as a one-to-many relationship between a Midi Track and several VST Instruments. Why do I want this you ask? Here is a recent example: I wanted to layer a Dim Pro Pad over a True Piano and would prefer this type of workflow:
1) Assign the output of a single Midi Track to Dim Pro and True Pianos,
2) Record the Midi Track.
3) Correct any wrong notes or poor timing in the single Midi Track.
4) Create a second Midi Track and copy data from the first into the second.
5) Remove the Dim Pro from the output of the first Midi Track and assign it to the output of the second.
6) Correct any issues specific to one instrument with velocity, etc on each track.
As you can see, Step 3 is much easier if there is only one track. Otherwise I need to keep copying my changes from track to track in order to preview my changes. Also, I often find that I don't need Steps 4-6 so the process is quite streamlined if we had layering as defined above.
Note: Suggestions for a better approach to this task within Sonar will be appreciated but it does seem like a one-to-many relationship between a Midi Track and several VST Instruments could be very useful.
Cheers,
R