hI

I'm building a drum machine sequencer abstraction (yep another one!) that will have 8 or so rows of 16 toggles. Each toggle has a unique send name (set in properties not [receive] object) and I'm writing to one very large table. Depending on which row is clicked, I want to send an number which can be used to multiply say 200 into the table. IE. CLick tgl in 2nd row -> 2 * 200 = 400 and start entering from there.

but I want to avoid having to create loads of [receive] objects for each toggle, for each row (potentially 128). Whats a clean dynamic way to do this so I can create a blueprint once and use this to enter into the table?

this will save me lots of time and will be easier to edit!! Is this something to do with [receive13] or [iem_receive]?

thanks

hmmmmmmmmmmmmm

[iem_receive] ~= [receive13] ~= [receive] (> 0.40) but it's no use because you have to set the receive value before you change the toggle, which defeats the point of it.

I thought [value] was going to help, but you can't dynamicaly change value and it whines "no method for float" if you try and set it from a send symbol.

There would be a trick to scan all the toggles if you could query a toggle without toggling it, but you cant.

I'm stumped.

one for the wise ones on the list I think.

can you link the toggles for each row in series, so that whichever one is toggled in the row sends an integer 1 to the end of the row?

I am currently working on a similar drum machine using an array of toggles, and am also daunted by the prospect of having to set up so many receive objects, as I also want to include toggles or radios for accent.

What information are you intending to send to the table?

I found myself in an similar situation, and I found the text editor to be my solution. I arranged the names of all my send/receive pairs on a grid, like for 256 sends, I make a 16x16. Then I choose names for the columns and rows of the grid-

columns : one two three four five six....
rows: ONE TWO THREE FOUR...

and then I make a single column of receive objects in a pd patch and save to disk, like this;

[oneONE] [oneTWO] [oneTHREE] [oneFOUR]....

Next, I make a folder and put 16 duplicates of the patch and label them 1-16. Then I do find/replace on each multiple in the folder. If you do this enough, it pays to write a perl/c script for it.

You can also start them out with $0- so you can dynamically expand and contract these s/r banks. And you can make them with as many dimensions as you want. I've got a patch with [r~ alephONEfirstALPHA] in it, but it doesn't do anything yet.

does this make sense?

I'm doing similar stuff as sunji when i have a patch that contains loads of same things (e.g. 64 receives). When i encounter this i create for instance a subpatch which handles a reasonable amount of instances (e.g. like this :

[r row1Step1] [r row1Step2] [r row1Step3]...

Then open this subpatch in the texteditor of your liking, replace the row1 by row2, save under another name, open in pd and copy/paste (aargh). But it all depends on what you're trying to do of course...

I think that is the best solution for now - create subpatches of receives and edit in text editor. Its actually not too bad one you get into it...

Still if someone finds a dynamic solution then that would be cooler. Pd needs some kind of global event listener, and with that info it could extract the number of the toggle toggled....

thanks for your help

here here, pd should have local, global and environment scopes. Who'll code it?

Puredata does have global read/write vars in the form of [value]. At the moment they are *only* accesible through the [value] object afaics. Perhaps feature requests to make [value] more..er valuable, are to have a declared variable writable from a send symbol, be typed, and be readable by a dynamic key/pointer.... more like datastructures

This would work like so:

Create an object [value float a float b symbol c]
Create a GUI object like a toggle with send symbol = a, a slider with send symbol = b etc

Have a [value] be dynamic

[bang( [set a(
| /
[value]
|
[print]

some of the stuff you guys have been talking about is a bit over my head, as I don't know anything about C programming, and Pd is the first time I've ventured into any sort of programming. So I'm sorry if I'm asking the obvious, but could you explain to me how to set up a grid in Text Editor, and arrange send and receive objects in that grid with row and column headers? I tried copying things to Text Editor but nothing happened. Also not sure how to save the stuff in text editor as a separate file?

Appreciate any help you can throw my way!

sure thing, brettb- when I say set up a grid, i don't mean in the text editor, but simply on a sheet of normal paper. Label the columns (1-16) and label the rows (ONE-SIXTEEN) and then derive your s~/r~ labels from the grid you have built;

1ONE 2ONE 3ONE 4ONE
1TWO 2TWO 3TWO
1TRI 2TRI 3TRI
1FOR

like so. As far as using a text editor;

all pd patches are saved as plain text files, bu with a .pd extension rather than a .txt. So if you open a text editor and open a pd file, you should see something like:

#N canvas 0 22 389 638 10;
#N canvas 410 32 788 776 pulsilator 0;
#N canvas 66 22 727 768 OverToneSECONDpatch 0;
#X obj 284 340 outlet~;
#X obj 361 73 tabread OverToneSECOND;
#X obj 338 118 pack;
#X obj 123 98 inlet;
#X text 105 79 pitch of the fundamental;
#X obj 361 97 / 127;
#N canvas 0 22 482 332 countto7 0;
#X floatatom 250 207 5 0 0 0 - - -;
#X obj 248 178 int;
#X obj 320 177 + 1;
#X msg 120 108 0;

notice the lines that say "OverToneSECONDpatch"? there's a pd patch I use that has "OverToneTHIRDpatch" all the way up to "OverToneSIXTEENTHpatch" and I just multiplied them by opening them in my text editor (i Highly recommend linux's Gedit, or smultron) and replacing all instances of 'SECOND' with 'THIRD', etc...

If you're text editor won't open a .pd extension, relabel it .txt, open it with the txt editor, and after find/replacing, change it back to .pd.

does this clarify?

yes, it certainly does! Thankyou very much for taking the time to explain it out Sunji.

I was trying to do something like this by changing the non-zero values of my toggles. Unfortunately, the zero value for each toggle is obviously zero, so I had unique note-ons (1, 2, 3 ... 32) for my drum sequencer but not unique note-offs (0, 0, 0 ... 0).

Is there any way to change the zero-value of a toggle (sounds silly I know)? Or is there another possible gui object we could use for these sequencers?

Regardless, Sunji's text-based solution seems the most elegant (and least laborious). I am just now programming my own drum sequencer, and I was really wanting some global control over my objects so that I could fabricate a series of toggles or receive objects with serialized names and so on.

Think of the god like power, the gigantic sequencers we could create (quickly) if we could instantiate 128 serially-named toggles and their corresponding receive objects!

I'm just musing now, but is it possible to create an algorithm in Pd that would create a textfile featuring these serialized objects? Something like an [until] loop that would create, for example, 32 (or 64, or 128) toggles, one at a time, featuring names built from symbols. I'm finding it hard to articulate, so maybe I'll just have to try to program it.

Thanks for the inspiring, relevant conversation.

D