Soft Synth MIDI and Audio Tracks - Managing tips?

Any tips on how to organize MIDI and AUDIO tracks for soft synths? 

For example, do you group all the MIDI tracks together, and all the Audio tracks together?  Or do you group each MIDI/Audio pair together?

Do you use the audio volume and pan, or the MIDI volume and pan?

How do you keep things from becoming confusing?

Sorry, by grouping I meant putting the tracks into track folders.

Does that mean you keep all your drums in 1 midi track?  Isn't that hard to edit?  Or do you use track layers?

Also, how would you setup folders for:

a) multiple midi inputs and single audio outputs (e.g., TTS-1 or Kontakt)?

b) multiple midi inputs and multiple audio outputs (e.g., Kontakt)?

I think the main benefit is being able to adjust the pan and volume of each drum sound independently.  Also, sometimes we'd like to use drum sounds from different kits.

Some soft synths don't respond reliably to volume and pan automation in the MIDI track. I have some, for example, that you can't completely silence by lowering an automation node to zero. So I tend to apply volume automation in the audio track. It also lets me raise or lower volume by specific db amounts, something that's hard to do via MIDI.

However, I prefer to automate swells in the MIDI track because it's easier to visualize when the automation is displayed over the MIDI data.

Not all soft synths pan stereo outputs correctly, so whenever I'm using a stereo out I insert an instance of the Channel Tools plugin on the audio track to pan it.

(Note, too, that most orchestral libraries are already panned according to the standard classical stage setup, so if you're after a natural-sounding orchestra no panning is required.)

I admit my process requires more housekeeping than other ways, but it helps keep me organized and sane.

I've seen one of Michael's projects, and if anybody has figured out how to stay organized in a big project, he's the guy.

There was a time when I would keep the drums in a single MIDI track until I was 99% done with the beats themselves and then split them out into separate tracks (kick, snare, toms and overheads) for audio processing. But I only did that to work around limitations of the synths I was using.

I don't do that anymore because the synths I use now don't require it. I can drive Superior Drummer from a single MIDI track and split out the audio any way I want within SD. My standard setup is one MIDI track and 5 audio tracks, gathered into a single folder. Any ancillary percussion will also go into the same folder for the sake of organization.

But that's just for drums. For orchestral stuff, I usually keep the MIDI and audio tracks in separate folders because I'm usually working on one or the other, not both. Once the synths have all been bounced or frozen, I'll archive and hide the MIDI folder to keep things tidy.

Funny...in all other aspects of my life I am neither organized nor tidy. My filing system is based on the geological principle of superposition, i.e. the oldest layers are on the bottom of the pile.

There are different ways we can route through a soft synth:

one-to-one = each MIDI track has its own AUDIO output track
many-to-one = multiple MIDI tracks share a single AUDIO output
one-to-many = a MIDI track gets turned into multiple AUDIO outs (e.g., drums)

Hardware synths typically result in a many-to-one setup.  On my Triton I had 16 MIDI tracks going to it, with 1 AUDIO out.  My projects consisted of 16 MIDI tracks and 1 AUDIO track.  All automation was done on the MIDI tracks.  This kept the Sonar side of things simple, but I had to manage effects on the Triton itself, which was cumbersome and limiting (I couldn't use any software effects).

Most people on the forum seem to be recommending a one-to-one setup (or even one-to-many in terms of drums).  This results in at least as many AUDIO tracks as MIDI tracks (unless you use Simple Instrument Tracks, which no one seems to recommend).  You then need to decide how to organize your MIDI and AUDIO tracks within your project.

You can either group-by-type (MIDI tracks together and AUDIO tracks together) or group-by-instrument (each instrument's MIDI and AUDIO tracks right next to each other).  Most people on the forum seem to recommend group-by-type.

Group-by-type encourages a "two phase" approach to song creation: MIDI first, AUDIO second.  During the MIDI phase you define the performances and arrangement, and during the AUDIO phase you mix.  If you bounce your MIDI tracks to audio, the AUDIO phase is the same as if you had recorded real instruments.  For people familiar (or wanting to be familiar) with the traditional studio environment, this seems to be the most natural way to work.

There are a few downsides to group-by-type:

1) You are discouraged from applying effects, panning, or other AUDIO track tweaks during the MIDI phase.  By restricting yourself to MIDI edits, you must either limit yourself to the synth's built-in effects (which means no Guitar Rig), or you must perform "dry" while trying to imagine how the instrument will sound with effects applied.  This clearly hinders inspiration.  My guess is that most people "cheat" a bit here.

2) [if you do not bounce] Since the AUDIO tracks are always empty (no clips) and the MIDI track is not nearby, it's hard to see when to apply your automation changes during the mix.  For people who always "mix with the ears" (e.g., console view) this might not be an issue, but for those of us who use our eyes as well (track view), this discourages precise mixing.  Note: It would be nice if Sonar could "mirror" a MIDI track within an audio track until you actually record audio.

3) [if you bounce] Bouncing consumes lots of disk space, unless you always delete the bounced tracks later.  (On the other hand, keeping your bounced tracks is the only way to guarantee that you can remix long after your plugins have expired or changed). Question: Is there a way to automatically delete bounced tracks?

4) [if you bounce] Bouncing discourages MIDI tweaks once you start mixing.  If you are in the middle of a mix and notice a weak note (or interaction between notes), you have to consider the hassle of unfreezing, editing, and re-freezing versus just letting it go (or, worse, trying to fix it in the mix). 


Experienced professionals will likely favor the one-to-one, group-by-type, bouncing approach, since it most closely matches the traditional studio environment.  Their essentially unlimited disk space and their good ears (for spotting problems in performance or arrangement before the mix), and good imagination (enabling them to perform dry) will mitigate most of the disadvantages I listed above.

Hobbyists are the ones who suffer most from those disadvantages.  Bouncing becomes expensive, but if you don't bounce, it gets harder to mix in audio, so you tend to mix in MIDI (no bouncing).  Performing dry is daunting, so you tweak effects before performing, which encourages you to move your audio tracks next to the corresponding MIDI tracks (group-by-instrument).  If you just want to keep things simple, you can even go for the many-to-one approach, which requires no bouncing or grouping.

It would be nice if the Simple Instrument Track worked well enough for professionals and hobbyists alike.  However, no one seems to advocate using it.  What are the main problems with it?

Also, I am curious: what approach do you use, and why?

I like your going by sections idea.

It's almost as if we have one synth per mic (or mic array).