Hi David and Christopher,
I am not so quick to dismiss the thoughtful list of characteristics
David submits here. It points the inexperienced ear in the right
direction. Christopher describes the gap between the two experiences
pretty much the way it occurs to me - worlds apart.
That said, there's something going on here in my own experience.
Midi playback has made me lazy about exercising my inner ear, so my
score reading is not what it should be. That's one side of it. The
other side is that I seem to have developed a kind of inner
translator, an imperfect but useful one, that gives me a decent idea
of what to expect from live musicians when I hear the "little kazoo
band". (Thanks, Andy Homzy, for that wonderful expression.)
Improvements in midi playback are always welcome, but even at its
perhaps primitive level of development, it serves its purpose for me.
I don't know how common an experience this is. I can be fooled by
some of the better examples (short ones, anyway) that I've heard
posted here, I think because my mind is responding to elements of
musical design that overwhelm my paying attention to the surface
information. If there are no glaring distractions, I am drawn to the
musical idea more strongly than I am to the limitations of the sounds.
I already know the sound of a saxophone, or a saxophone section, and
I am disappointed by midi reproduction of that. (I always know
immediately when I'm hearing midi saxes.) But my mind translates it
pretty well, and I am able to tell if I have written what I want,
even with the limitations of the midi sax sounds. (Only the bari
comes close enough to recognize as a sax sound in Garritan JABB, and
it's not because Gary and Tom haven't worked on it, or because they
started with bad samples. It has to be something in the nature of
the saxophone sound envelope as played by live musicians that is so
different and has so much more variety than what the machine creates.)
The brass sounds aren't so hot either - but they do their jobs for me
in providing a useful sketch, and the percussion sounds, including
guitar, piano and bass, seem pretty good to my ear.
What is never right - in my scores, are the balances. Those things
change so rapidly in a live performance, that I can't even imagine
the amount of work it would take to include them in sound created by
a machine, and whatever you created that way would change in the next
live performance. So I mistrust the midi playback for this and apply
my experience with live bands to judge whether or not I have a chance
of having written what I want.
Darcy has made thoughtful suggestions to Garritan for improvements in
the balances, things that would get them started with a closer
approximation of what you'd hear a band play, but those things vary
so much from band to band that I'm not sure there is a "right" way to
set up the midi balances. Orchestral balances are more standardized
- maybe as a result of having refined them over the long history of
orchestral music and the constant adjustments that have been made in
numbers of players in the sections and the traditions of orchestral
playing.
We have a tougher time in jazz bands. The traditional
instrumentation is dynamically out of whack to begin with, and few
bands compensate for this in a way that would satisfy my musical
vision. (It is possible to do - check out The Sultans of Swing
recordings for a contemporary example, or the Basie, Ellington and
Mulligan bands for historical standards of balances that work.)
So what I want to suggest to users of this technology is something
like: Check your pitches and rhythms, with the understanding that
subtleties in each of those areas will be changed by live musicians
(in good and bad ways!), and try to get an idea of the overall
textural and timbral (is that a word?) structure - the similarities
and contrasts. Then imagine that most of them will be vastly
improved in any halfway decent live performance.
Hows that for a cautionary instruction?
Thanks for the help in thinking about this.
Chuck
On Jun 23, 2007, at 7:26 AM, Christopher Smith wrote:
On Jun 23, 2007, at 6:55 AM, dhbailey wrote:
[snip]> I am working on an interactive jazz arranging book/DVD -
whatever it's
going to turn out to be - for Gary Garritan. I am convinced that
this technology provides a useful sketch tool for composer
arrangers who don't have musicians available, but it only serves
as a severely limited interim sketch tool.
I wish I were better equipped to describe the differences between
midi instrument playback (at the mundane level of notation
software playback) and living musicians. It might be a useful
descriptive skill to include in the book.
Chuck
[snip]
These comparative remarks assume that the midi playback will
always be through the same midi device, whether soft-synth or
hardware module:
1) midi instrument playback will always be the same, always
predictable, always have hits exactly the same volume in exactly
the same spots. Live musicians will vary their performance with
each new performance of the song.
2) midi instrument playback will not interact with the live
performer at all, so that if a human musician gets some new
rhythmic groove going it won't matter one bit to the midi
instrument playback, there will be no variation. Live musicians
will follow and build on such a lead, so that a real dialog will
get going.
3) midi instrument playback will always demonstrate the same
"envelope" around the sound, same attack, same release, same
vibrato at the same rate and same depth every single time the song
is played. Live musicians will alter all aspects of the sound,
sometimes consciously just for variety's sake, sometimes
unconsciously as their bodies change from day to day.
4) midi instrument playback will always have the same tone with no
atmospheric or personality variations from day to day or season to
season. Human musicians can't demonstrate quite the same
consistency, since certain physical attributes of their
instruments vary as weather varies (humid/dry, cold/hot).
5) midi instrument playback, by definition, involves listening
through and amplifier and speakers. These never replicate the
acoustic properties of hearing human musicians playing acoustic
instruments. When the humans are playing only electronic
instruments such as electronic drums, electric piano, electric
guitar, this difference becomes much less noticeable, but when
humans are playing on acoustic instruments, the physics of sound
are quite different and the listening sensation is very different
than hearing even the same sounds through speakers.
6) human aspects such as personality, good-day/bad-day, happy,
angry, moody, sad, and other psychological factors will never
become part of midi instrument playback, yet they have an enormous
effect on the sound of human playback.
Heh, heh!
Seems like an excellent start, but I think Chuck was thinking along
the lines of "here's a recording of a MIDI band playing the
passage" and "now here's a recording of the same passage played by
live musicians."
Just that step is a huge one. I often have a "Finale kazoo band"
playback and a gorgeously recorded CD of the same piece to play for
clients so that they recognize what they are listening to when I
create a mockup for them, and so that their imaginations can make
the necessary leap.
Most of my students do their projects on Finale or Sibelius, and
sometimes they ask me if they can submit an MP3 of the playback
instead of preparing parts and bringing them to the reading
session. I always respond with an emphatic "NO", as a huge part
(perhaps the biggest part!) of composing and arranging for live
musicians is hearing what the live musicians do with your piece.
Besides, I can read the score pretty well, so the playback isn't
for ME anyway!
Christopher
_______________________________________________
Finale mailing list
Finale@shsu.edu
http://lists.shsu.edu/mailman/listinfo/finale
Chuck Israels
230 North Garden Terrace
Bellingham, WA 98225-5836
phone (360) 671-3402
fax (360) 676-6055
www.chuckisraels.com
_______________________________________________
Finale mailing list
Finale@shsu.edu
http://lists.shsu.edu/mailman/listinfo/finale
