Subject: MAX Digest - 28 Sep 1998 to 29 Sep 1998 (#1998-16)
Date: Wed, 30 Sep 1998 00:00:01 -0400
From: Automatic digest processor 
Reply-To: MAX - Interactive Music/Multimedia Standard Environments
     
To: Recipients of MAX digests 

There are 15 messages totalling 545 lines in this issue.

Topics of the day:

  1. navigation
  2. video
  3. a Karplus-Strong petcs baja... (2)
  4. block size / feedback
  5. X<>Pose
  6. AiffPlayer (2)
  7. MAX Digest - 27 Sep 1998 to 28 Sep 1998 (#1998-15) (2)
  8. random and standalone app.
  9. high notes
 10. Query re: MIDI->infrared
 11. Num object to chr. & Variable base
 12. fiddle~ for MSP

Email to MAX should now be sent to MAX@lists.mcgill.ca
LISTSERV commands should be sent to listserv@lists.mcgill.ca
Information is available on the WEB at http://www.mcgill.ca/cc/listserv

----------------------------------------------------------------------

Date:    Tue, 29 Sep 1998 01:05:47 -0400
From:    Stephen Kay 
Subject: navigation

>ryohros
>NAVIGATING WITHOUT A MOUSE:
>I can't use patchcords because of too many numberboxes . is there any
>transparent
>easier solution?

Can it be possible that you do not know that patch cords can be made
invisible with Command-K? (Hide On Lock)  Perhaps you do not have a
manual?  Or perhaps your question is not clear.

>EFFITIENCY CONSERVATION:
>Does graphic faders,using send & receive and the overuse of numberboxes
>can slow down the overal preformance of the machine?

The larger your patch and the processing it does, then obviously
the performance may suffer.  Use number boxes where you need them.  Do
not leave thousands of them hooked up in every particular place for
debugging.  Yes, use them in the debugging stage.  Afterwards, get rid
of most of them - you can always put them back later.  True, a numbox
should not do anything when it is hidden - but it still must receive
the message and *decide* not to do anything.  On a slow machine with
1000s of numboxes, it could add up.

What machine are you using?

Stephen Kay

------------------------------

Date:    Tue, 29 Sep 1998 07:36:05 +0200
From:    Jeffrey Burns 
Subject: video

>> I've come to rely on X<>Pose to display full screen video and PICTs with
>>Max.
>> It works very well. The drawbacks I find are (1) X<>Pose doesn't support
>> hardware compression, which means I've had to use cinepak (resulting in
an
>> appreciable drop in image quality), and (2) X<>Pose doesn't allow random
>> access to the video, but always starts at the beginning.
>
>I'm planning a project which will trigger video from Max. Would you mind
>giving a brief explanation as to the advantage of using X<>Pose? Why not
>just use the QuickTime movie object of Max? It allows random access and
>supports hardware compression with the Miromotion 30 and G3 Buz, so I've
>heard. What are the drawbacks of the Max object? I may be in a position
>to create an updated version of it in a few months, so please give your
>suggestions.

Is it really true that X<>Pose handles video better than Max, given that
the same compression is used in both programs? If so, it would be good if
Max's video speed could be raised to the level of X<>Pose, as it's stupid
to run 2 programs simultaneously, just because one is a little faster at
playing cuties. It would also be interesting if the Max movie object could
do something with alpha channels - although Image/ine is the program for
doing heavy video processing.

BTW, does anybody have experience with Max and a firewire video board?

Jeff Burns

http://www.snafu.de/~jeff

------------------------------

Date:    Mon, 28 Sep 1998 23:00:32 -0800
From:    dudas 
Subject: a Karplus-Strong petcs baja...

Georg Hajdu writes:
>I've been playing around with some of the CNMAT night school patches
>recently and was very intrigued by the Karplus-Strong patch. Unfortunately,
>it wasn't possible to create pitches higher than middle c (due to the fact
>that tapin~/tapout~ can't handle delays shorter than 1.5 msec). I replaced
>tapin~/tapout~ by the delay~ object (which is suitable for shorter delays)
>and noticed that--while delay~ worked fine with simple delays--is wasn't
>possible to do feedback. I simply had no output.

Whereas it seems logical to program a feedback loop in the way that you
showed in your patch, it just doesn't work. If you think about it, you are
telling the signal compiler to add an object's output to its own input
before it has even had the chance to calculate the output signal! (so the
DSP doesn't even run... you've created an infinite loop.) Try doing the
same kind of connection between an int object and a number box and you
should get a lovely stack overflow and have to restart the max scheduler.

A sneaky way to do feedback in your MSP patch would be to add a send~ and a
receive~ on either side of the delay~. Since the send~/receive~ combination
makes a copy of the signal buffer, the signal compiler won't run into the
infinite loop problem.  However your signal will be delayed by the signal
vector size (it seems you're using 64 samples), and thus you get the same
pitch restriction as with the tapin~/tapout~. (Back to square one, it
seems...)

So, what you need is an obejct that has a sample-by sample delay with
built-in feedback... i.e. the comb~ object.  (In fact, comb~ without
feedback does the same thing as delay~.) Les Stuck did a superb example
called "Kazoo" which can be found in the MSP examples folder.

Actually, it might be nice to have a karplus object with a built-in biquad~
filter and 4-point interpolation for the delay. Anyone want to give it a
try???

Remeelem hogy az segiit, vagy legalaabb vilaagiit egy kicsit mieert nem
muekoedik a petched. Lehet hogy probaalok iirni egy karplusz obdzsektet
valamikor az eev veege elloet, ha igazaan haszaalnos lett volna neked.

-R

------------------------------

Date:    Tue, 29 Sep 1998 00:32:18 -0600
From:    Kevin Walker 
Subject: Re: a Karplus-Strong petcs baja...

>Actually, it might be nice to have a karplus object with a built-in biquad~
>filter and 4-point interpolation for the delay. Anyone want to give it a
>try???

I have a Karplus-Strong object which allows for small delays.  It also
allows for a variety of "plucking" techniques (plucking with formants is
fun).  I've been meaning to make in available for months -- as soon as the
help patches are cleaned up a bit -- but other things have take priority.
If anyone's interested, I can make the current messy version available.

Kevin

------------------------------

Date:    Tue, 29 Sep 1998 11:16:03 +0200
From:    Robert Henke 
Subject: block size / feedback

Hi maxers,

(...) > I've been playing around with some of the CNMAT night school =
patches
> recently and was very intrigued by the Karplus-Strong patch. =
Unfortunately,
> it wasn't possible to create pitches higher than middle c (due to the =
fact
> that tapin~/tapout~ can't handle delays shorter than 1.5 msec). I =
replaced
> tapin~/tapout~ by the delay~ object (which is suitable for shorter =
delays)
> and noticed that--while delay~ worked fine with simple delays--is wasn't
> possible to do feedback. I simply had no output. Even a simple patch =
with
> feedback such as the following didn't work (...)

this is definitly one of the hardest msp cons.  i was told that its a bit
tricky to handle feedback in this kind of software, but as far as i know =
it
is possible : the "generator" software from native instruments can do it
and is basically the same: combining dsp objects in a graphical
environment. it also has a block size of 1. I have the impression that (
due to the block size? ) its more cpu consumtive than msp but at least it
is possible to do things like coupled spring - mass oscillators , karplus
strong etc...
neverthesless i would prefer to do this all in max/msp !!!
david, will it be possible in the future ?

rob.

------------------------------

Date:    Tue, 29 Sep 1998 09:41:23 +0100
From:    "David Z. Saltz" 
Subject: Re: X<>Pose

> Going back to cinepak sounds like a sad story. Does anybody know if
> hardware
> compression support it's on the way?

About a year ago, someone at Steinberg told me that a "pro" version of
X<>Pose
was on its way with support for various flavors of hardware compression. I
haven't heard anything about it since, and I don't know if they're still
planning to release such a product.

------------------------------

Date:    Tue, 29 Sep 1998 09:57:48 +0100
From:    "David Z. Saltz" 
Subject: AiffPlayer

Over the past couple of years, I periodically download the latest version of
Eric Singer's Aiffplay/er object and use it in a project. Unfortunately,
each
time I end up having to replace it with the snd object, and often end up
sacrificing some of my patcher's functionality as a result. The problem is
simply stability: Aiffplayer brings down my system after 10-20 minutes of
steady use.

I've tried increasing the buffer size, varying the number of voices, nothing
helps. Over the years I've experienced this problem on a wide variety of
machines, including PowerMac 8100s, G3s and Powerbooks, running various
versions of Max.

I've been reluctant to bring this problem up, because the object isn't an
official part of Max, and it is only to the good graces of Eric Singer that
it
exists at all. But I really, really like the object, which is why I keep
trying to use it. From scattered comments on the list over the years, I get
the sense that lots of people use this object. Has anyone else had similar
problems? Is there a solution?

------------------------------

Date:    Tue, 29 Sep 1998 16:26:01 +0200
From:    Roland Cahen & Ruth Sefton-Green 
Subject: Re: MAX Digest - 27 Sep 1998 to 28 Sep 1998 (#1998-15)

Thanks to
Roby Steinmetzer 
& Matthew Ostrowski 

This is NumToList*
an object based on Roby's to tranform a number into a equivalent list
of variable lenght in any basis
it is the answer to both questions (Num object to chr. & Variable base)

max v2;
#N vpatcher 230 58 720 451;
#N comlet List Output;
#P outlet 28 329 15 0;
#P number 354 72 35 9 0 0 0 3;
#P newex 354 90 35 196617 int $2;
#N comlet Basis;
#P inlet 354 53 15 0;
#N comlet Lenght of the output list;
#P inlet 310 53 15 0;
#P newex 310 90 35 196617 int $1;
#P number 310 72 35 9 0 0 0 3;
#P number 167 71 35 9 0 0 0 3;
#P newex 167 89 27 196617 int;
#N comlet Model value input;
#P inlet 167 52 15 0;
#P number 245 126 43 9 0 0 16 3;
#P number 52 126 110 9 0 0 2048 3;
#P number 310 162 35 9 0 0 0 3;
#P number 354 209 35 9 0 0 0 3;
#P newex 52 166 56 196617 t b i;
#B color 1;
#P newex 52 216 56 196617 int;
#B color 5;
#P number 167 126 70 9 0 0 0 3;
#P newex 22 246 27 196617 / 2;
#B color 5;
#P newex 52 246 27 196617 % 2;
#B color 5;
#P newex 52 327 60 196617 prepend set;
#B color 9;
#P newex 52 279 84 196617 group 16;
#B color 9;
#P message 52 351 189 196617 0 0 8 13;
#P newex 52 303 44 196617 reverse;
#B color 9;
#P newex 52 191 40 196617 Uzi 16;
#B color 1;
#P comment 392 210 32 196617 Basis;
#P comment 349 162 115 196617 Lenght of the output list;
#P comment 52 111 63 196617 Binary input;
#P comment 166 111 43 196617 Decimal;
#P comment 245 111 30 196617 Hexa;
#P comment 151 250 100 196617 Roby Steinmetzer ;
#P comment 151 277 164 196617 + Roland Cahen
;
#P comment 3 32 156 196632 NumToList*;
#P comment 4 65 112 196617 Converts a Number into a variable lengh list
written in any basis;
#P comment 151 304 109 196617 Thanks to Matthew Ostrowski
;
#P connect 19 0 10 0;
#P connect 19 1 18 1;
#P fasten 18 0 16 0 57 242 27 242;
#P connect 18 0 15 0;
#P fasten 16 0 18 1 27 267 18 267 18 212 103 212;
#P connect 15 0 13 0;
#P fasten 23 0 19 0 250 154 57 154;
#P connect 14 0 12 0;
#P fasten 17 0 19 0 172 154 57 154;
#P connect 13 0 11 0;
#P fasten 11 0 33 0 57 323 33 323;
#P connect 11 0 14 0;
#P connect 10 0 18 0;
#P fasten 20 0 16 1 359 242 44 242;
#P fasten 20 0 15 1 359 242 74 242;
#P connect 29 0 27 0;
#P connect 24 0 26 0;
#P connect 30 0 32 0;
#P fasten 22 0 19 0 57 154 57 154;
#P fasten 21 0 10 1 315 190 87 190;
#P fasten 21 0 13 1 315 190 131 190;
#P connect 26 0 25 0;
#P connect 25 0 17 0;
#P connect 28 0 21 0;
#P connect 27 0 28 0;
#P connect 31 0 20 0;
#P connect 32 0 31 0;
#P pop;

Yours Sincerly
Roland Cahen
http://www.cicv.fr/CICV/TB/ARTISTES/rolcahen.html
http://www.gmeb.fr/LesCompos/Lescompos97.html#anchor6344194
MandalaElectronique téléchargeable à :
ftp://ftp.ircam.fr/pub/forumnet/max/FAT/applications/
http://www.cicv.fr/IC/

------------------------------

Date:    Tue, 29 Sep 1998 17:10:10 +0000
From:    Jakob Brandt-Pedersen 
Subject: random and standalone app.

Hello.

I am working on a cdrom-project, that use a Max standalone application.

When I make a standalone whith random objects, they are not random in
the application. This is because random get it seed from the time since
system startup and it is of course always the same.
I have tried to work around this by loading the patch into another patch
at a later time, but that doesn't help.
Does anyone have any suggestions how to get randomination into
standalone applications?

Jakob Brandt-Pedersen.
Multimedia-artist.

jbrandtp@get2net.dk
http://hjem.get2net.dk/jbrandtp/

------------------------------

Date:    Tue, 29 Sep 1998 12:32:51 -0400
From:    Eric Singer 
Subject: Re: AiffPlayer

>Over the past couple of years, I periodically download the latest version
of
>Eric Singer's Aiffplay/er object and use it in a project. Unfortunately,
each
>time I end up having to replace it with the snd object, and often end up
>sacrificing some of my patcher's functionality as a result. The problem is
>simply stability: Aiffplayer brings down my system after 10-20 minutes of
>steady use.
>
>I've tried increasing the buffer size, varying the number of voices,
nothing
>helps. Over the years I've experienced this problem on a wide variety of
>machines, including PowerMac 8100s, G3s and Powerbooks, running various
>versions of Max.
>
>I've been reluctant to bring this problem up, because the object isn't an
>official part of Max, and it is only to the good graces of Eric Singer that
it
>exists at all. But I really, really like the object, which is why I keep
>trying to use it. From scattered comments on the list over the years, I get
>the sense that lots of people use this object. Has anyone else had similar
>problems? Is there a solution?

I'd like to know this too.  In fact, I'd like to hear bug reports on any of
my objects, as I am happy to try to address them, time willing.  But I can
only do this if I get user feedback.  So, users, don't feel reluctant to
post problems with non-official objects - this is the forum to do this.

David, are you using the latest version I have on my site
(http://ericsinger.com)?

-Eric Singer

------------------------------

Date:    Tue, 29 Sep 1998 09:33:51 -0700
From:    David Zicarelli 
Subject: high notes

Georg Hajdu  writes:

>I've been playing around with some of the CNMAT night school patches
>recently and was very intrigued by the Karplus-Strong patch. Unfortunately,
>it wasn't possible to create pitches higher than middle c (due to the fact
>that tapin~/tapout~ can't handle delays shorter than 1.5 msec). I replaced
>tapin~/tapout~ by the delay~ object (which is suitable for shorter delays)
>and noticed that--while delay~ worked fine with simple delays--is wasn't
>possible to do feedback. I simply had no output. Even a simple patch with
>feedback such as the following didn't work

It's not possible to do feedback with delay~, MSP detects the
feedback loop in the signal path and refuses to make it work.
With tapin~ and tapout~ the connection between the objects
is made with a control connection, so you can do feedback.

The minimum signal vector size will be reduced to two samples in the next
major release of MSP, so people who want to do high Karplus-Strong
notes will be a lot happier. Why two and not one? I don't know, one didn't
work too well and I figured two was small enough. Then I can get
Georg to pay for another upgrade when I fix it in a few years. See,
you too can be in the software business in five easy lessons.

David Z.

------------------------------

Date:    Tue, 29 Sep 1998 12:07:37 -0700
From:    Ken Gregory 
Subject: Re: MAX Digest - 27 Sep 1998 to 28 Sep 1998 (#1998-15)

Hello Jeff,

We 've been using the Max imovie QuickTime object with the Targa 1000 video
card from TrueVision on a G3 with a SCSI Fast n'Wide hard drive controller.
Our initial tests were with a PPC 9500. The cool thing about the Targa card
is that it outputs the movie seperately from the desktop. No fussin with
the desktop and Max patches to make it to look right. Also the Targa has
it's own compression method that allows fullscreen full resolution 30 FPS
video playback recording and still look good. ALso any random access of
your movie with your Max patch  is quick and snappy.

I've also done a live audio/video mix performance with the Targa 2000
running on an 8500 with a Fast n'Wide SCSi hard disc controller. No
problems with random access, video scratch, looping, looping on the fly and
all that other fun stuff.

Hope this helps.

Cheers,
ken gregory

>
>
>> It works very well. The drawbacks I find are (1) X<>Pose doesn't support
>> hardware compression, which means I've had to use cinepak (resulting in
an
>> appreciable drop in image quality), and (2) X<>Pose doesn't allow random
>> access to the video, but always starts at the beginning.
>
>I'm planning a project which will trigger video from Max. Would you mind
>giving a brief explanation as to the advantage of using X<>Pose? Why not
>just use the QuickTime movie object of Max? It allows random access and
>supports hardware compression with the Miromotion 30 and G3 Buz, so I've
>heard. What are the drawbacks of the Max object? I may be in a position
>to create an updated version of it in a few months, so please give your
>suggestions.

Ken Gregory     kgregory@pangea.ca
704-138 Portage Ave. East
Winnipeg, Manitoba, Canada R3C 0A1
ph. 204-942-7510  fax 204-942-1555

------------------------------

Date:    Tue, 29 Sep 1998 12:27:57 -0700
From:    Martin Gotfrit 
Subject: Query re: MIDI->infrared

My students are often interested in controlling video decks (or
televisions, or ..etc.) via Max.To avoid taking the thing apart and hard
wiring a relay it seems to me there must be a MIDI to infrared or serial to
infrared gizmo out there (perhaps for presentation or teaching
applications).  Any ideas?  Thanks in advance.
Martin.

_________________________________________________________
Martin Gotfrit * Associate Professor, Music Area
School for the Contemporary Arts, Simon Fraser University
Burnaby, B.C. Canada  V5A 1S6
* 604 291 3517 * fax 291 5907 *

------------------------------

Date:    Tue, 29 Sep 1998 23:16:44 +0100
From:    Roby Steinmetzer 
Subject: Re: Num object to chr. & Variable base

>This is NumToList*

Now it would be nice if we could determine the lenght of the output list en
fonction du nombre input.
I tried sending the input through an expr object with some logarithm stuff.
Though this works well with log10, I have some problems with other bases.
These can be due to my lack of mathematical understandings but I'm sure
some bright mind on the list has a solution to this.

Roby Steinmetzer
Luxembourg, Europe

------------------------------

Date:    Tue, 29 Sep 1998 17:57:43 -0800
From:    Ted Apel 
Subject: fiddle~ for MSP

Yes, just in time for Miller's talk on fiddle~ and bonk~ at the ICMC. The
first MSP versions are available at
http://www-crca.ucsd.edu/~tapel/software.html

Please send bug reports to me at tapel@ucsd.edu

Thanks,
Ted Apel

>From:    Jeff Mann 
>Subject: fiddle for MSP?
>
>Hi - someone posted a while back that he was working on a port of Miller
>Puckette's "fiddle" voice-tracking object, for MSP. Can't remember who
>it was... any progress?

------------------------------

End of MAX Digest - 28 Sep 1998 to 29 Sep 1998 (#1998-16)
*********************************************************