Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site sdcarl.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!whuxlm!harpo!decvax!ittvax!dcdwest!sdcsvax!sdcc3!sdcarl!andy
From: andy@sdcarl.UUCP (Andrew Voelkel)
Newsgroups: net.music,net.music.synth
Subject: Re: Drum Machines - A Flame
Message-ID: <195@sdcarl.UUCP>
Date: Thu, 30-May-85 15:05:14 EDT
Article-I.D.: sdcarl.195
Posted: Thu May 30 15:05:14 1985
Date-Received: Sat, 1-Jun-85 23:25:45 EDT
References: <317@mhuxr.UUCP> <979@pyuxd.UUCP> <320@mhuxr.UUCP> <988@pyuxd.UUCP> <327@mhuxr.UUCP> <992@pyuxd.UUCP> <328@mhuxr.UUCP>
Reply-To: andy@sdcarl.UUCP (Andrew Voelkel)
Organization: Computer Audio Research Lab, UCSD, San Diego, Calif.
Lines: 34
Xref: watmath net.music:7744 net.music.synth:296

>pastoral mood that day. It is a measure of Max's creativity that
>he completely altered the rhythmic feel of the piece, without
>changing the beat. THAT is what a drummer can do. Not every one
>is Max Roach, but any decent drummer will adapt similarly to a change
>in melodic intensity. As I see it, a preprogrammed machine, almost
>by definition, cannot react to an unfolding development. The
>melodist thus has not rhythmic challenge; Time deadens, and stops
>altogether. Result:  dead music.
>

I don't want to get into this argument, but I would like to propose that
a drum machine need not by definition be "followed", which is what everyone
(myself included) has been doing to date. The marvelous thing about
midi is that more complex control arrangements can be designed on a 
separate computer that can use just the timbres of a drum machine,sampler
or synth. Given the proper input devices (keyboards, "simmons" pads w/midi,
guitar interfaces), the electronic drum set that Marcel accepts as a
valid instrument (am i wrong?) is but a trivial subset of the possibilities.
Preprogrammed parts could follow live performers, or interact with them
in more interesting ways. Such improvisational algorithms fascinate 
me, and I see that such tools designed properly could actually open up
new possibilities for interaction between human performers because some`
of the actual sounds can become "shared resources".

I have been laying the groundwork for developing these ideas by writing
a public domain device driver for the Roland MPU-401 and the ibm pc.
So far I have just the normal record and play functions working, but
the architecture will be open so as to encourage other musically mot-
ivated programmers to provide the utilities I wont (ie fancy graphics
or fancy editing). I'm almost ready to release what I've got. 

p.s. The reason for using the Roland box is that four can be ganged 
together for multiple realtime midi inputs. This is to provide for
multiple "live" performers