[
Lists Home |
Date Index |
Thread Index
]
Something like that, but MIDI is sequencing and sequencing
of that type depends on a controller issuing a sequence
to and through a set of channels where each channel has
the instructions (ADSR, note, etc.) that any instrument
can listen to and respond to, but that the orchestrator
will segregate (drums are almost always on channel 10),
but the rest are up for grabs. In other words,
you may have multiple performers. That let's you do
interesting things like layering a sound (instruction
parallelism).
The problem is latency (in/out) and thru or the
saying "sometimes the speed of light isn't fast
enough.
If the devices are hooked by midi in/out/thru of
the performing devices, the aggregation
will produce a noticeable response lag after approx.
three devices. A box is used to get the signal and
distribute it simultaneously to all devices, so
the controller sends midi out to the distributor
and it sends midi out to all members of a performance.
Also, some devices can receive midi in, but can't pass
midi out or thru. Thru is used if you want the device
to pass the instruction without reacting to it.
A pipeline isn't the ultimate form of orchestration
if you really need the 'ineluctable modality' of
parallelism. That is why orchestras have conductors
who hold batons and musicians learn to watch both
the baton and the individual copy of the score in
front of them. It is why sequencers have thru
and the midi protocol has assignable channels.
It is also why coordinated systems often need
a signal handler for distributing channel information.
Welcome to Hytime and why it was interesting to
enterprise designers even though it started out
as a music notation language. Remember, the modality
of time scales (absolute beat to division representation
to baton movement). The latency of eye movement and
breathing gives it "feeling" as well as the conductor's
movements. Strict midi feels mechanical. Strict
orchestration of enterprise processes do as well.
Style counts.
The comparison to REST is in the instruction types,
that is, what information does a midi instruction
contain and how to extend it. General MIDI is
typical. Beyond that, it gets very task and even
device specific.
But orchestration middleware has a future. Not
every device is a J.S. Bach, capable of improvising
in six independent voices from a theme.
... dang... now I have to go back to working
on the Fugue In G minor for lute and labor
over a tool-inappropriate task. I think I'll
just do Shakira's Ojos Asi instead. It is
more shocking and less work.
len
-----Original Message-----
From: Sean McGrath [mailto:sean.mcgrath@propylon.com]
At 09:55 01/03/2002 -0600, Bullard, Claude L (Len) wrote:
>Pipelining is fine. Someday they will discover MIDI. :-)
All software development is a footnote to J.S. Bach.
Conversations between independent voices conducted in accordance
with the sole ineluctable modality - Time.
|
