Subject: MAX Digest - 17 Aug 1999 to 18 Aug 1999 (#1999-246)
Date: Thu, 19 Aug 1999 00:00:16 -0400
Automatic digest processor <LISTSERV@LISTS.MCGILL.CA>
Reply-To: MAX - Interactive Music/Multimedia Standard Environments <MAX@LISTS.MCGILL.CA>
To: Recipients of MAX digests <MAX@LISTS.MCGILL.CA>

There are 11 messages totalling 430 lines in this issue.

Topics of the day:

  1. loading files into mtr
  2. about snd files
  3. Realtime FX processor
  4. sfcontrol~?
  5. CPU and pcontrol problems
  6. Computer Opera "Der Sprung" in M=?ISO-8859-1?B?/A==?= nster
  7. dynamic level to MIDI data?
  8. Pluggouts
  9. dac~ breathing
  10. groove~ bug?
  11. video Sampling: MAX vs. software


Date:Tue, 17 Aug 1999 20:24:24 -0500
From:Fred Collopy <flc2@PO.CWRU.EDU>
Subject: loading files into mtr


Is anyone aware of how I might automatically load a file into mtr (without
going through the standard mac load dialogue)?

thanks, Fred

Fred Collopy
Assoc Prof at Case Western Reserve University and
Occassional Visiting Scientist at IBM's T.J. Watson Research Center


Date:Tue, 17 Aug 1999 20:47:50 -0400
From:Otto Henry <muhenry@EASTNET.EDUC.ECU.EDU>
Subject: about snd files

Alexander H. asked anbout snd files-

I always have problems with complex snd file names- sometimes no matter
Max won't see them unless I just name them "fred" or something simple
hope this helps,


Date:Tue, 17 Aug 1999 17:40:30 -0700
From:Mark Dumas <trailerstudios@WORLDNET.ATT.NET>
Subject: Re: Realtime FX processor

>I'd like to use my sequencer (i use Logic) to process my voice in

>realtime with various plugs (standard VST and expecially Pluggos).

>But i've just discovered that, even if all meters show my live input,
>nothing goes through plugs and i just hear my dry voice :( it
>Is there a way to use Logic (or any other audio/seq, apart
>MAX/MSP!!!) as a realtime FX processor?
>I hope someone has a solution...

SPARK is a two track audio editor that accepts multiple VST plugins that
will act as an effects processor. From the manual p.36:

"By choosing 'Realtime FXmachine' from the Options menu, you can turn
your computer into a Realtime FXmachine. This will route your sound
input's signal to the FXmachine and return the processed signal to the
output. This way you can use your FXmachine without recording


Date:Wed, 18 Aug 1999 00:14:54 -0400
From:Jeremy Bernstein <jeremy@BOOTSQUAD.COM>
Subject: sfcontrol~?

At 12:00 AM -0400 8/18/99, Automatic digest processor wrote:
>There are many ways to do this with Max/MSP
>try putting your sfiles in a single folder in the max folder
>(if sfile name not reckognized, put the folder in the pref path)
>use sfplay~ or better sfcontrol~ with two voices.
>write your sfnames in as many messages boxes you want
>trig one box or another with a select object controled by velocity
>or whatever.
>That could be a start
>try also using buffer~ and groove~ or other ram sound disposalls for short
>Good luck
>Roland Cahen

sfcontrol~? what's that?


- - - - -
"The collage technique ... is the most important innovation
in the art of this century. Found objects, chance creations,
ready-mades ... abolish the separation between art and life.
The commonplace is miraculous if rightly seen, if recognized."
-- Charles Simic, Dime-Store Alchemy: The Art of Joseph Cornell.

current project: _1999: An Insomniac's Walking Surveillance Journal_
is updated daily at

PGP public key available from


Date:Wed, 18 Aug 1999 15:39:53 +0200
Subject: Re: CPU and pcontrol problems


>From:John Phillips <jpch@TERRAGIZMO.NET>
>Subject: CPU and pcontrol problems

>I'm working on an MSP patch with 8 patchers of "sound-makers" heading for
>one dac~ -all with a toggle box hooked to enable $1 then into
>outside of the patcher. I use a coll to select which modules are "on." My
>beginner's question is: when all the patchers are "off" and I toggle the
>dac~ on, my CPU usage is around 18-20%... when several of the patchers are
>on it only goes up a percentage point or two. I've checked by looking into
>the patchers that are "off" and can't see any activity. Maybe just a dac~
>breathing uses that much CPU? UMAX/G3 board, OS 8.1, plenty of RAM.

I assume you use a faster machine. I'm more or less afraid, that enabling
or disabling the patchers don't make much difference. I'm working on a
7100/80, which is really slow compared to anyone with a 604 or G3 machine.
On this machine everything is much more extreme, so I might say something
what I ASSUME what might happen.

First, I made a larger patch with several abstractions. My idea was to have
one module active at a time with a short crossfade between two when
changing the active module (so for a short period of time two modules are
active). I made a general moduleframe containing the enable/disable logic
and the abstraction of the specific module. (The name of the module was an
argument to the moduleframe instance. I made a #1 abstraction with the
correct number of inlets and outlets for connecting the module instance in
the moduleframe.) Basic test showed that it worked. Even the selection of
different modules show a changing CPU load. After that, I asssumed, that it
wouldn't be a problem to make similar modules like the first one. BUT THIS
WAS NOT TRUE! I had to throw off Miller Puckette's fiddle~, a 256-point
fft/ifft and was still not able to build all eight modules. even
deactivating all patchers signal processing didn't give the CPU time back.
For a long time I didn't understand, why this could be possible. Now I have
two explanations:

a) The program doesn't fit into the cache (my cache is only 256K). Deactivated
modules should be thrown out of the cache automatically after a short time,
but this doesn't happen. The reason for might be found in b).

b) If signal processing is deactivated, it means, that the calculation load
goes down, but it does not mean, the patcher an inactive program. It rather
means, that only the event processing goes on. But I'm afraid that the
processing and signal processing objects are rather small and interleaved,
so that the signal objects couldn't be thrown out of the cache without the
event processing objects (probably signal processing objects do both, event
and signal processing and couldn't be thrown out of the cache for that
too). Unfortunately I didn't try to deactivate the event processing
part too, I'm rather busy preparing for a concert and things must run.
Hopefully I might try this next month...

So for your problem, It might bem that your cache is full and so the load
goes up without much calculating going on. The real difference in
calculating might not be very much compared to the slowdown which results
by using the slow main memory. Just a guess. I hope David Z. may shed some
light on this if he got some time for that...




Date:Wed, 18 Aug 1999 18:03:43 +0200
From:Georg Hajdu <hajdu@UNI-MUENSTER.DE>
Subject: Computer Opera "Der Sprung" in M=?ISO-8859-1?B?/A==?= nster


I am proud to announce the world premier of my computer opera "Der Sprung" =
after a libretto by Berlin author and filmmaker Thomas Brasch.

"Der Sprung" will be performed in M=FCnster, Germany at 8 pm on October 2nd
and 3rd, 1999 at the "Theater im Pumpenhaus" in a concert version.

The opera is based on a real story: It's about madness, a killing, a German
university, the German-Jewish problem and a high-diver who is about to do
her final, deleterious leap.

On my web site, I documented the making of the opera which was derived from
a single recorded statement by the librettist. Departing from this
statement, Brasch and I formed virtually every aspect of the opera from it,
although occasionally in a very free manner.

For the composition, I have used dozens of MAX programs, including neural
networks and patches that enabled me to do automatic, non-standard,
transcriptions into Finale (microtonal music and space notation).

"Der Sprung" is scored for narrator, 8 singers ("wired choir"), 3 electroni=
and 17 acoustic/electric instruments. Its middle part, Intermezzo, won 1st
prize at the Braunschweig synthesizer composition contest and honorary
mention at the Blaue Br=FCcke competition in Dresden.

Please visit the web site at m=
homepage at further information.


Ich m=F6chte hiermit die Urauff=FChrung meiner Computer-Oper "Der Sprung" nach
einem Libretto des Berliner Schriftstellers und Filmemachers Thomas Brasch

"Der Sprung" wird in einer konzertanten Fassung in M=FCnster am 2. und 3.
Oktober 1999 um 20 Uhr im Theater im Pumpenhaus aufgef=FChrt werden.

Die Oper beruht auf einer wahren Begebenheit: Es dreht sich um Wahnsinn,
t=F6dliche Sch=FCsse, eine deutsche Universit=E4t, das Deutsch-J=FCdische Problem
und um eine Turmspringerin, die ihren letzten verderblichen Sprung tut.

Auf meiner Webseite habe ich die Entstehung der Oper dokumentiert, die aus
einem einzigen aufgenommenen Satz des Librettisten abgeleitet wurde.
Ausgehend von diesem Satz haben Brasch und ich so ziemlich jeden Aspekt der
Oper entwickelt, gelegentlich jedoch in sehr freier Weise.

W=E4hrend des Kompositionsprozesse habe ich Dutzende von MAX-Programme
geschrieben, darunter solche, die neuronale Netzwerke zum Einsatz brachten
oder die es mir erm=F6glichten, mikrotonale Musik automatisch in Finale zu

Der Sprung hat folgende Besetzung: Sprecherin, 8 S=E4nger, 3 elektronische un=
17 akustische/elektrische Instrumente. Der mittlere Teil, Intermezzo, wurde
bereits mit dem 1. Preis der Neuen Akademie Braunschweig f=FCr
Synthesizerkomposition sowie mit einer Auff=FChrungs- und F=F6rderungsempfehlun=
der Blauen Br=FCcke, Dresden ausgezeichnet.

Ich m=F6chte Sie einladen, meine WebSeite unter meine Homepage unter besuchen.

Ich hoffe, Sie im Oktober in M=FCnster begr=FCssen zu d=FCrfen.

Dr. Georg Hajdu

Hochschule f=FCr Musik Detmold,
Abteilung M=FCnster
Ludgeriplatz 1
D-48151 Muenster



Date:Wed, 18 Aug 1999 12:42:00 -0500
From:chriso <chriso@RIPCO.COM>
Subject: dynamic level to MIDI data?

Newbie question:

Is it possible to create a patch that would convert (in realtime) audio
levels to MIDI data? Is it possible to analyze a realtime audio waveform's
frequency, and create MIDI from this?



Date:Wed, 18 Aug 1999 14:28:58 -0400
From:Jeff Brown <jeffb@AMNH.ORG>
Subject: Pluggouts

> Date:Tue, 17 Aug 1999 15:34:01 +0100
> From:
Roland Cahen & Ruth Sefton-Green <>
> Subject: Pluggouts
> Why not doing 'pluggouts' instead ?
> that meens separate application
> I did some with MSP and truly it's much more fun.
> I which someone would develop some good interface
> which would be able to run any pluggins
> including pluggo and commercial vst pluggins.

What? This sounds like an idea I'd be interested in, if I ... err ...
understood it.


Date:Wed, 18 Aug 1999 14:38:19 -0400
From:Jeff Brown <jeffb@AMNH.ORG>
Subject: dac~ breathing

---John Phillips wrote:
> question is: when all the patchers are "off" and I
> toggle the
> dac~ on, my CPU usage is around 18-20%... when several of the
> patchers are
> on it only goes up a percentage point or two. I've checked by
> looking into
> the patchers that are "off" and can't see any activity. Maybe
> just a dac~
> breathing uses that much CPU?

What's your CPU usage when you turn off the dac~? A lot of that CPU
usage is probably unrelated to the specific patch, but overhead just for
using MAX in general, and other system stuff. I know *my* machine's
like that . . .



Date:Wed, 18 Aug 1999 23:14:10 +0200
From:Gilbert Nouno <gilbert.nouno@FREE.FR>
Subject: Re: groove~ bug?

well, as for me, I think groove~ 's OK .

I findit nice to access sample level rather than decimal ms,
and you can always use the mstosamps~or sampstoms~ to do the conversion.

Concerning the sync output,
I think the whole real length of the sample is scaled from 0. to 1.,
but of course it's not the lenght of the loop, so unless the loop is all
the sample, it is scaled between begin > 0. andend < 1.

The sync signal reaches 1. to tell that it has completed the loop.
This is the thingBen was considering as a bug when saying sync
doesn't go from 0. to 1. within a loop.
I must say I find this groove~behaviour very cleaver,
because, if you put an " ==~ 1."followed with "edge~",
you get a bang each time the loop is starting again, moreover
you always know where you are in absolute time in the sample and in the loop !

let' say my loop is scale between 0.2 and 0.6 (in respect
to the whole sample),
then if you see the sync output (with print~),
you will have

0.2 ..... 0.5.... 0.58 .. 0.59 .. 0.6... 1. 0.2 ..... etc.

hope it helps, anybody correct me if I'm wrong !


>Date:Tue, 17 Aug 1999 10:44:59 +0100
Benjamin Thigpen <Benjamin.Thigpen@IRCAM.FR>
>Subject: Re: groove~ bug?
>Hi Johan,
>Thanks again for your response.
>>>I wonder if this is behind my other problem - setting the loop points of
>>>groove~ with signals. No one responded to that message, so I suppose it
>>>works fine for everyone else ????
>>Actually it did not with me (using latest msp version). Indeed groove~
>>accepts signal as samples, rather then milliseconds. Also the sync output
>>is not updated with a signal as loop-end.
>Ah. This is bad news. I did not yet manage to update (file not found
>error), but I did replace pow~ and groove~. Although replacing pow~ cured
>my other problem (thanks), replacing groove~ did nothing. Seems to be a
>bug then, no?
>>Question (though): what would be de advantage of using signals for

so that you can change them without clicks. If you're using the
>sync signal to put an envelope on your loop (fadein, fadeout), you can use
>sah~ to change your loopoints only when the envelope is at 0. Otherwise
>you get clicks:
if you change the looppoints mid-loop, it changes your
>sync signal instantaneously and so jumps to another point in your envelope


Date:Wed, 18 Aug 1999 18:33:13 -0500
From:chriso <chriso@RIPCO.COM>
Subject: video Sampling: MAX vs. software

I would like commentary on the difference between video sampling
software(i.e. vidvox, image/ine, xpose, etc.), and building a quicktime
playback patch from the ground up in MAX.
In terms of access speed for a quicktime clip- Is a MAX patch
comparable in speed tocustom C code for video sampling?



End of MAX Digest - 17 Aug 1999 to 18 Aug 1999 (#1999-246)