Laccum + normalize with "forget" rate?

Apr 14, 2006 at 10:47am

Laccum + normalize with "forget" rate?

Hello All,

I’m trying to build a list accumulator with normalization. I’ve got that far (though I’m severly mathematically challenged, so please feel free to fix what I already think I’ve done!), but what I want to add is an optional “forget” rate. What I’d like is a slightly more stubborn version of what I already have — one that will change more slowly over time. Make sense?
[ edit: that is, same peak of 1.0, but smoother curve... yes? ]

Here’s what I’ve committed, thus far (need Lobjects, btw):

#P window setfont “Sans Serif” 9.;
#P window linecount 1;
#P newex 148 135 45 196617 t l l;
#P flonum 208 179 35 9 0 0 0 3 0 0 0 221 221 221 222 222 222 0 0 0;
#P newex 148 200 70 196617 Ldiv;
#P newex 183 158 35 196617 Ltop 1;
#P newex 148 245 46 196617 t l clear;
#P newex 82 280 76 196617 Laccum;
#N comlet lists in;
#P inlet 82 26 15 0;
#N comlet normalized accumulated output;
#P outlet 82 357 15 0;
#P connect 6 0 5 1;
#P connect 4 1 6 0;
#P connect 7 1 4 0;
#P connect 3 0 2 1;
#P fasten 5 0 3 0 153 229 153 229;
#P connect 7 0 5 0;
#P fasten 2 0 7 0 87 307 251 307 251 111 153 111;
#P connect 2 0 0 0;
#P connect 1 0 2 0;
#P fasten 3 1 2 0 189 271 87 271;
#P window clipboard copycount 8;

#25423
Apr 14, 2006 at 11:40am

On 14-Apr-2006, at 12:47, jbmaxwell wrote:
> I’m trying to build a list accumulator with normalization. I’ve got
> that far (though I’m severly mathematically challenged, so please
> feel free to fix what I already think I’ve done!), but what I want
> to add is an optional “forget” rate. What I’d like is a slightly
> more stubborn version of what I already have — one that will
> change more slowly over time. Make sense?

If the list size is constant you could do your “forget rate” easily
with a bank of lp.stacey objects.

Here’s an example for 3-element lists and a forget rate of 5.

#P window setfont “Sans Serif” 9.;
#P window linecount 1;
#P newex 58 96 216 196617 pack 0. 0. 0.;
#P newex 225 70 92 196617 lp.stacey 5;
#P newex 122 70 92 196617 lp.stacey 5;
#P newex 19 70 92 196617 lp.stacey 5;
#P newex 19 47 217 196617 unpack 0. 0. 0.;
#P inlet 19 24 15 0;
#P outlet 58 121 15 0;
#P connect 6 0 0 0;
#P connect 5 3 6 2;
#P connect 4 3 6 1;
#P connect 3 3 6 0;
#P connect 1 0 2 0;
#P connect 2 0 3 0;
#P connect 2 1 4 0;
#P connect 2 2 5 0;
#P window clipboard copycount 7;

If you require variable list lengths or changing forget rates you
would need to do some scripting with this approach. At least stacey
does the math for you.

Just an idea.

– Peter

————– http://www.bek.no/~pcastine/Litter/ ————-
Peter Castine +—> Litter Power & Litter Bundle for Jitter
Heavy-Duty Mathematics for Everyday Use
iCE: Sequencing, Recording &
Interface Building for |home | chez nous|
Max/MSP Extremely cool |bei uns | i nostri|
http://www.dspaudio.com/ http://www.castine.de

#74861
Apr 14, 2006 at 12:40pm

Thanks Peter,

I didn’t realize stacey had that option, but alas, my current application uses a list of around 50 members (I’m doing some fuzzy stuff), so it would be a little cumbersome to do with stacey. I think I can get my head around the math, so I may do it in java…

btw, I had a look at dspaudio and became quite curious about iCE. Is there any more documentation I can look at to get a better idea of how it might be used, or what it’s capable of doing? I could just use a little more detail.

thanks,

J.

#74862
Apr 14, 2006 at 1:52pm

On 14-Apr-2006, at 14:40, jbmaxwell wrote:
> btw, I had a look at dspaudio and became quite curious about iCE.
> Is there any more documentation I can look at to get a better idea
> of how it might be used, or what it’s capable of doing?

We’re working on it. As it becomes available it will be on the
DSPAudio site.

Best, Peter

————– http://www.bek.no/~pcastine/Litter/ ————-
Peter Castine +–> Litter Power & Litter Bundle for Jitter

iCE: Sequencing, Recording & |home | chez nous|
Interface Building for |bei uns | i nostri|
Max/MSP Extremely cool http://www.castine.de

http://www.dspaudio.com/

#74863

You must be logged in to reply to this topic.