On Tue, 28 Sep 2004 10:40:25 -0400, Dave Phillips <dlphilp(a)bright.net> wrote:
A good question. I've transcribed a lot of piano
music by various
composers, and I enter each note from the computer keyboard. I follow
the dynamics indicators in the music, then I apply a very little
humanization to start-times, durations, and velocities. After that I do
edit a tempo track manually. Again I follow the tempo indicators in the
music, but I also provide an underlying tempo loop (usually at the level
of the 16th note or 8th note triplet) to "disturb" the even playback
typical of step-entered MIDI music. My sequencer of choice has a very
easily edited tempo track, so all this stuff takes less time than you
might think.
I might give that approach a try; some of the stuff I want to get into
MIDI format is over my head in terms of what I can perform without
massive practice. :)
I find that with some careful shaping of dynamics
and tempo you can
create the illusion of a decent performance. Of course, you have to know
what a decent performance is, and of course you need to know something
about the performance practice for the kind of music you're
transcribing, but all this works well for me.
If you were recording a piece that you know well enough to perform (if
you're not a pianist, pretend your instrument of choice (even voice)
were a midi input device), would you be as satisfied?
This has got me thinking. I think the ideal for me would be to record
free and then mold the tempo track to the recording. Rosegarden will
let you do a beat track, perhaps I'll try that. But what if you could
just drag the barlines (or beats, or your choice of granularity) to
the appropriate spots?
--
De gustibus non disputandum est.