Jeff,
At the heart of this discussion is a dilemma that really has no practical solution at present. Anyone composing in a notation program must realize its strengths and weaknesses for MIDI applications. Working in a notation program has the strength of being familiar and comfortable, where ingrained writing habits needn't be modified greatly from good old paper and pencil. The finished product can be ready for publishing, printing, and presentation to musicians for live performance. Also, MIDI playback of the parts can be set up to give you some feedback on the sound of your work as you proceed. The direct notation data can also be "massaged" to give a more musical result than the straight quantized data of the notation, if you like. As I said, it can give *more* musical results but rarely results to equal playing the parts one at a time into a sequencer. That's the strength of sequencing - much more believable results can be achieved if the primary concern is the success of the final mix. Playing the parts into a sequencer has little to do with being "keyboard indulgent" and everything to do with expression. Actually, each part should be played in from the perspective of the instrument being simulated. You would not play a solo violin part with the same approach as a solo flute part (and neither would be played with keyboard technique in mind). This, of course, requires you to be intimately familiar with the idiosyncrasies of each of the instruments to be simulated. Just as in an ensemble of real musicians, playing each of the parts in separately adds small "human" inaccuracies and interpretive tendencies for timing, note length, etc. that can make for a much more expressive and convincing result - it's easier to bring a mix "alive." Editing additional detail (attacks, volume shaping, breathing spaces, etc.) into these basic performance tracks can add even greater nuance and believability. But these characteristics rarely translate into good notational results. Usually they translate into awful notational results.
It all comes down to this: If the music is intended primarily for live performance use the notation program and don't be overly concerned with the MIDI mix results. But (now comes the dilemma): If you want to work in notation and also have the best possible mix, one way or another you're stuck with considerably more work. You can take the printed notation and re-record every part into a sequencer to get the killer mix or you can take the great sequenced mix and edit the data to be excellent notation. There is really no way around this at the present time. You just need to make the choice that is best for you.
Tom
At the heart of this discussion is a dilemma that really has no practical solution at present. Anyone composing in a notation program must realize its strengths and weaknesses for MIDI applications. Working in a notation program has the strength of being familiar and comfortable, where ingrained writing habits needn't be modified greatly from good old paper and pencil. The finished product can be ready for publishing, printing, and presentation to musicians for live performance. Also, MIDI playback of the parts can be set up to give you some feedback on the sound of your work as you proceed. The direct notation data can also be "massaged" to give a more musical result than the straight quantized data of the notation, if you like. As I said, it can give *more* musical results but rarely results to equal playing the parts one at a time into a sequencer. That's the strength of sequencing - much more believable results can be achieved if the primary concern is the success of the final mix. Playing the parts into a sequencer has little to do with being "keyboard indulgent" and everything to do with expression. Actually, each part should be played in from the perspective of the instrument being simulated. You would not play a solo violin part with the same approach as a solo flute part (and neither would be played with keyboard technique in mind). This, of course, requires you to be intimately familiar with the idiosyncrasies of each of the instruments to be simulated. Just as in an ensemble of real musicians, playing each of the parts in separately adds small "human" inaccuracies and interpretive tendencies for timing, note length, etc. that can make for a much more expressive and convincing result - it's easier to bring a mix "alive." Editing additional detail (attacks, volume shaping, breathing spaces, etc.) into these basic performance tracks can add even greater nuance and believability. But these characteristics rarely translate into good notational results. Usually they translate into awful notational results.
It all comes down to this: If the music is intended primarily for live performance use the notation program and don't be overly concerned with the MIDI mix results. But (now comes the dilemma): If you want to work in notation and also have the best possible mix, one way or another you're stuck with considerably more work. You can take the printed notation and re-record every part into a sequencer to get the killer mix or you can take the great sequenced mix and edit the data to be excellent notation. There is really no way around this at the present time. You just need to make the choice that is best for you.
Tom