Hmmm!
A good question.
And no, no flowchart, at least not as yet.
Let me just put down briefly my own approach.
But, before I do, let me add that for me 99% of the time, I'm creating classical music backing tracks, much of which is done from notation.
I use XGworks almost exclusively in the early stages, and so my description follows what is allowed in that sequencer. I only move to Sonar 7 PE for the later process of rendering the MIDI to audio and thence to CD.
Initially I sort out the 'master track'. This holds all the meta events such as tempo, time and key signature, copyright, sequence name and all that.
These all go at 1:1:000.
I set the first measure's tempo (120 also at 1:1:000) and the tempo to be used for the music in measure 2:1:00 along with time and key signatures. If the music undergoes a change of key or time signature part way through, then I work out where this will be and enter it, e.g. at 22:1:000. The same with major changes of tempo.
Of course entering and changing key signatures is only essential if you work in staff view, but is good practice otherwise.
What I often do, especially now that I primarily use virtual instruments, is to keep the the first measure (the set up measure) very short. I reckon a second is good, which is one measure of 2/4 time. So the first measure is set to 2/4, the second to the time sig. of the music.
If you're sending SysEx setup messages to a synth at the start of a song, or many changes to the patches using RPNs and NRPNs, then often a bar of 4/4 time is usually adequate, but be prepared that it may need more.
Once I've done this I then move to the first track, setting the Bank Select MSB and LSB, then Program Change messages in the first measure, I do a copy and paste job of all the controllers I might use from a previous song, setting them to default or 'unused' values if I'm not sure what exact levels they may need to be. So, e.g. channel volume, pan, reverb send and so on.
The various Garritan orchestral, band and jazz libraries that I use, use a lot of MIDI controllers for all sorts of control of the sound of the instruments, so I may have up to 20 controllers entered with their start up values. They then get modified as I go along entering the notes.
This procedure is then duplicated for the next track, and so on.
Once all the notes for all the tracks are in, I go back and enter some variation to note velocity and timing to give a more human feel.
Then back to the start and work on dynamics. Then back again and work on intonation. Then phrasing, then levels of vibrato and so on.
My aim is to get each track sounding correct as a solo performance before trying to get a good mix.
That's about it off the top of my head.
Any use?
JohnG.
