Best Practices for MIDI tracks?

I mistakenly posted some questions on the wrong forum.  They belong here.
 
My intention is to have one audio track for each MIDI track (or possibly multiple audio tracks for one MIDI track in the case of drums.)  That being the case, it seems to me that I should plan to do all the volume, panning, and FX on the audio tracks.  In other words, I should set the MIDI fader at unity and dead center on the pan.  And I should not mess with any mixer settings inside the synth VSTs.  Is that what y'all would consider best practices?
 
If I am going to keep all my original MIDI in the project (rather than bouncing to audio), is it best to put all the MIDI tracks in their own folder and then just collapse the folder so I never see the lights and never touch the MIDI again (unless going back to revise the source material.)?

I sometimes have pan, reverb, and chorus commands in my MIDI tracks while I am arranging, but I remove them before recording and process everything on the audio track as you say.  I don't think the fader does anything to a MIDI track, but I could be wrong.
 
EXTERNAL SYNTHS:
I may have a mutli-timbral synth (like a Motif or something) coming into one audio track while I am arranging.  For example, I may have six or seven instruments coming into Sonar in stereo from one synth. Then, at some point when I am happy with the MIDI tracks, I do this:
1. Remove the effects and panning from the MIDI tracks (unless the synth has some unusually good digital effects and I am sure I don't need a dry version)
2. Clone the audio track several times.
3. Mute all but one MIDI track and one audio track and record the synth onto the audio track.  Once a track is recorded, I turn its input and echo off so it plays only the recorded audio and not the other MIDI tracks.
4. Now I mute, the MIDI track I already recorded, and repeate the process until I have recorded all the separate instruments to individual tracks.
5. After recording, be sure all the MIDI tracks are muted so they don't play anymore.
6.  Note that before recording, you may have to set the Controller 7 value appropriately in the MIDI track to get the audio at the right level.
 
SOFT SYNTHS:
I generally use pan and volume in the audio track; however, some soft synths have very good effects and I may use them.  For example, I use the reverb and EQ built into Synthogy's Ivory II.
 
I put all MIDI tracks in a MIDI folder, mute the foilder, and hide the folder after I am finished with them so they don't confuse me during the mix.

you don't really need to set up 1 for 1 doing drums.
here youll see @ 0:45 http://www.youtube.com/watch?v=TjZS61Wg-XY
I have all midi tracks for each drum piece but I can still add effects into the bins almost like its an audio track.
 
(unless of course there is a reason you need them to be audio  ;)

dmbaer

konradh
I don't think the fader does anything to a MIDI track, but I could be wrong.
 

I believe it maps to MIDI volume.  However, there's no guarantee that a soft synth will react accordingly.  I too could be wrong, but I'm under the impression that MIDI volume was an early standard that's largely gone by the wayside.

I just did a test with TTS.  When I push the MIDI fader up, it gets louder.  If I pull the fader down to zero, there is no sound.  The velocity indicators (green lights) are unrelated to the fader -- they keep going regardless of the fader setting.


I haven't tried that with every synth I have, but it has worked that way ever time I tried it, hence my question about best practices.  It sounds like the best practice is to leave the MIDI fader at the default level (101?).

I do tend to leave the midi alone... no panning of volume control in there..... But I'm pretty sure it can be done.... I know I have used the midi volume fader to raise the levels once or twice.
 
There is no problem with doing it that way, nor is there any problem with tweeking the controls in a synth or FX to get the desired result. What matters most is not how you get there but what it sounds like when you get to the final mix.
 
As far as best practices? Whatever you decide to do and how you decide to work, using whatever method and workflow that works best for you .... is what you should do. There are no hard and fast rules.... well not many. And rules are meant to be broken.
 
As far as what to do with midi tracks after the fact. I use midi tracks on occasion but once I have them right, I bounce them to audio and prefer to work on the final mix using all audio tracks where possible. So once the bounce is done, I drag the midi track and the synth track associated with it, to the bottom of the track view and archive it. That shuts the synth off. Then I hide the tracks...or perhaps I will even delete the synth track totally. If it's a piano, for example, I know I can pop it back in if I need it again. I do the same thing in the console view.... hide them or at the very least make them narrow.
 
When working with audio tracks I will do the same thing..... dragging them to the bottom of the track view if they are not active (muted) in the project. Doing this makes for a cleaner work space on the screen. I really don't like to hide tracks in both the console and track view because I tend to forget they are there..... so muting to the bottom works best. I will hide in the console view.
 
Hope this helps.

cparmerlee
I mistakenly posted some questions on the wrong forum.  They belong here.
 
My intention is to have one audio track for each MIDI track (or possibly multiple audio tracks for one MIDI track in the case of drums.)  That being the case, it seems to me that I should plan to do all the volume, panning, and FX on the audio tracks.  In other words, I should set the MIDI fader at unity and dead center on the pan.  And I should not mess with any mixer settings inside the synth VSTs.  Is that what y'all would consider best practices?

 
In my world it's the one and only practice! :) I like to have everything as un-preprocessed as possible and do all the sound shaping in Sonar's console. By doing this you always have everything that's going on right in front of you and never have to open any plugins to look for the effect (or what everI you want to change. I always do the mixing with the audio tracks and never touch volume or anything on the MIDI-tracks. There's nothing wrong with doing it any other way though. This is just how I work.
 
When I mix (which I do in CV) I've all MIDI-tracks hidden so just the tracks I need are visible. If I use a multi-timbral instrument I always use several outputs to get each channel on a separate audio track.
 

If I am going to keep all my original MIDI in the project (rather than bouncing to audio), is it best to put all the MIDI tracks in their own folder and then just collapse the folder so I never see the lights and never touch the MIDI again (unless going back to revise the source material.)?

Just as I want to have all the sound shaping tools available right in front of me when I mix I also want to have all MIDI-tracks in front of me when I do the MIDI. Therefore I always have one folder with all MIDI tracks in it. I hate looking for things so it's great to always have everything you need right there.
 
There's nothing wrong with bouncing to audio, I think. With the freeze-function in Sonar you don't loose the MIDI information by doing this. It's just to jump back and forth with one mouse click on the snow flake. :)
 
The audio tracks (both from software instruments and recorded audio) I place in different folders, one for each instrument category so to speak. I posted this video in another thread onces, but here you can see the folder structure I use: http://www.youtube.com/watch?v=EfMw6Q1HoIY

M_Glenn_M
This is good. But I'm not following the extraction process.
So how do I get all the different (drum) instruments that are in one (TTS1) midi track and make an audio track for each that I can tweak?
I've been just freezing and getting one audio track.

I don't think you can within TTS.  That is going to do all the drums on one track.  But Session Drummer can split each kit piece out to a separate track.

It is possible to use a CAL to put each note on a separate track. I am not at my music computer so I can’t look up the exact CAL name but I believe it is something like separate by note. As you know each drum is on a different note thus you will have separate tracks for each drum. You can then bounce to track each track individually.
I hope this helps and good luck.

So the "Split Note To Tracks.CAL" would be good for only the 4 in TTS1?
To get more we would use SD or other plugs right?

Guitarhacker
I do tend to leave the midi alone... no panning of volume control in there..... But I'm pretty sure it can be done.... I know I have used the midi volume fader to raise the levels once or twice.
 
There is no problem with doing it that way, nor is there any problem with tweeking the controls in a synth or FX to get the desired result. What matters most is not how you get there but what it sounds like when you get to the final mix.
 
As far as best practices? Whatever you decide to do and how you decide to work, using whatever method and workflow that works best for you .... is what you should do. There are no hard and fast rules.... well not many. And rules are meant to be broken.
 
As far as what to do with midi tracks after the fact. I use midi tracks on occasion but once I have them right, I bounce them to audio and prefer to work on the final mix using all audio tracks where possible. So once the bounce is done, I drag the midi track and the synth track associated with it, to the bottom of the track view and archive it. That shuts the synth off. Then I hide the tracks...or perhaps I will even delete the synth track totally. If it's a piano, for example, I know I can pop it back in if I need it again. I do the same thing in the console view.... hide them or at the very least make them narrow.
 
When working with audio tracks I will do the same thing..... dragging them to the bottom of the track view if they are not active (muted) in the project. Doing this makes for a cleaner work space on the screen. I really don't like to hide tracks in both the console and track view because I tend to forget they are there..... so muting to the bottom works best. I will hide in the console view.
 
Hope this helps.

Since there is no audio out of a MIDI track, does the MIDI volume control the velocity of the synt, therefore making it seem louder?
 
Walt

I think I'm right in saying that you can use CC11 to vary the volume of a held Midi note as it's playing.
 
Velocity of course will apply to the whole note no matter how long it's held, or it's length.
 
So where does CC7 figure into all this? 
 
I need to brush up on my Midi theory.........

For most synths, CC7 and CC1 do essentially the same thing.  The concept was to use CC11 to vary volume within phrases and songs, and CC7 to set the overall level.  Some synths implement CC11 in odd ways, but I haven't come across one yet (that I remember) that does not use CC7 for master volume.

The MIDI "gain" knob in the Inspector is the Velocity offset slider in the MIDI track
http://www.cakewalk.com/Documentation/default.aspx?Doc=SONAR%20X2&Lang=EN&Req=Views.31.html
Just to add confusion the doc refers to Velocity Offset as "The change in velocity (volume) that will be applied to notes in the track on playback" (emphasis is mine)
 

Bristol_Jonesey
So what would be the "correct" technique for programming, say, a violin performance (within an orchestral context)?
 
I've been ignoring CC7 altogether, then using different velocities in conjunction with CC11 to create the ebb & flow of a real performance.

Use cc11 for dynamics, cc7 for setting gain.

konradh
For deeply sampled libraries, the velocity will also switch to different samples.

Exactly right.  And this isn't just true of samplers.  A lot of classic synth patches moduate filter cutoff with velocity as well as loudness.  So velocity cannot be counted on to regulate just loudness.  And of course, there was a time when synth keyboards didn't recognize velocity at all, so patches trying to duplicate those vintage sounds might not even recognize velocity at all.  So again, no guarantees that velocity controls loudness.  It's all a matter of the programming of any particular patch/preset/whatever-you-want-to-call-them.