Here's how I've managed to send and receive from a parent(multi instance) abstraction to multiple instances of nested sub abstractions.

within main abstraction [mainAb]
some data named $0var send it like this

initialize sub abstractions
pass names X_1 and X_2 in so you can send and receive data to these abstractions individually
[abX X_1 $0] [abX X_2 $0]

within sub abstraction abX initialized with X_1
$1 = X_1
$2 = mainAb's $0

for all instances of abX
to receive from main's
[r $2var]

to to send and receive locally and have mainAb access
within abX
within mainAb [r $0X_1varZ] and [r $0X_2varZ]

a way to get global data to sub abstractions of abX
within abX initialize [abY $2var]
within abY $1var = mainAb $0var

This information is useful say if you have multiple voices and want to have them share the same control

so say $0var is a send on a filter frequency slider
when you move the slider all instances of abX receive the new filter frequency

when you want information to or from a specific instance of abX say like you want to plot each of the voices outputs on a scope
use [s~ $2$1varZ] in abX
then in main [r~ $0X_1varZ] and [r $0X_2varZ] then you can write those to a table so you can see which voice is sounding

Now I have some questions.

number 1
how do you get your last operation not to be calculated first?
I had made this intricate tuning calculation for a Karplus - Strong 12 string.
It's microtuneable to non octave tunings so in order to detune the second strings it has to be done on the fly. So every now and then the thing plays way out of tune.
I put prints in enough places so I could see what was going on.
Well final operation on calculating the frequency was getting calculated first. Or at least it was getting printed first. So I started tracing trigger bangs and floats all the way back to the first bang that started the calculation. That was crazy While making the thing I had followed the path very carefully down but since some operations depended on other operations to get done first the bangs split in multiple directions. Instead of trying to get more confused in the tangled mess I started replacing

• , [+ ], etc.. with [expr] which made it a lot easier to comprehend still sometimes the thing plays out of tune.
  This leads to my next question. The Karplus-Strong patch I started with sets a [block~ 1] is it possible this small block size is causing some of the calculation to get interrupted.

If the things that need to be calculated are not done within one block does that screw things up?
I thought the control operations were only sent once per block anyways.
and is that so if you change the block size to anything other than 64?

And another question is it more processor intensive to use [expr] instead of

• , [+], etc..?

Can a moog like filter be done without writing c code? Where can I learn about designing ladder filters in pure data? I like all the ext moog filters I've heard I'm just wanting to keep my patches vanilla unless I write the external. Propellorhead's Thor Ladder filter does not sound like any of the moog filters I've heard in pure data. Maybe it has something to do with the frequency snapped to the harmonic series or something but I've done that with the pure data moog filters and don't get the same results. Maybe the Propellorheads are upsampling to do the filters.

Another weird thing is I made a bypass for my filters. So I had 6 voices, 6 oscillators per voice, 2 filters per voice so that is 36 oscillators and 12 filters running. I added a 3rd filter per voice and when they were bypassed my machine went up to 90% cpu usage and more. So I removed the 3rd filter. Still CPU usage would go up when the filters were bypassed so I added a switch to turn them off whenever they were bypassed that fixed the problem with that. I probably ought to switch the oscilators off when they are not in use as well so I can add more. Some of them are feedback oscilators and generate chaotic signals so I guess there need to be an option to leave them on.
Now the next question and that will do it for today.

how do you print your entire patch to a ghost script printer?
I tried scaling the page so I could see the patch on the screen and still only get part of the patch in the ps file.

How do I do modulation routing with a matrix and do it polyphonically?
I want anything that has an input to be able to be routed to anything with an output and be scaled along the way. My voices are allready huge and so I would like it all to be done from within the main patch. At what point does your patches get too big? oops.

>>number 1
how do you get your last operation not to be calculated first?...the bangs split in multiple directions<<

are you using triggers to split your bangs? ie, [t b b b]
that way, the execution order will always be defined, and can be traced. if you just take multiple signals from a single object, without a trigger splitting it, the order will be unclear.

hope someone can help with your other questions!

mod, thanks
I do use [t b b] and [t f b] and all variations needed.
That is what was wrong. I had to add an [f ] or 2 to pause the flow and ...

I had the output of a [t f f] crossed. haha! After a complete redesign and tracing the bang flow many many times I finally found it.

@billystiltner said:

And another question is it more processor intensive to use [expr] instead of

• , [+], etc..?
  Only negligibly so.

Can a moog like filter be done without writing c code? Where can I learn about designing ladder filters in pure data? I like all the ext moog filters I've heard I'm just wanting to keep my patches vanilla unless I write the external.

I haven't tried it myself, but with what I have read about Moog ladder filters, it would require upsampling due to the non-linearities of the distortion. Also, I'm not so sure it would be easy to do with Pd's raw filters. If that's the case, you'd have to get down and dirty with [block~ 1], single-sample delays, and/or [fexpr~] within this upsampled patch. In other words, it might be a cpu intensive pain in the ass.

Why must you only use your own externals? That's not exactly vanilla, either.

Another weird thing is I made a bypass for my filters. So I had 6 voices, 6 oscillators per voice, 2 filters per voice so that is 36 oscillators and 12 filters running. I added a 3rd filter per voice and when they were bypassed my machine went up to 90% cpu usage and more. ...<snip>...

Sounds like a denormal issue. Try sending very quiet noise through them at all times.

Which filters are you referring to, btw?

At what point does your patches get too big?

They don't. Your computer just gets too small.

@Maelstorm said:

Why must you only use your own externals? That's not exactly vanilla, either.

I anticipate compiling my externals for use with mac, microsoft, and linux. I guess I should have said vanilla except for my own externals. There are some things in PDX that would make things so much nicer like the graphical envelopes, and image, but do all of these work well on all operating systems?

Is anyone aware that midiout on windows is broken? song position pointer and continue do not work. This was tested not with the latest version of PD but an earlier version. start and stop do work though as well as midi clock.

@Maelstorm said:

Which filters are you referring to, btw?

The filters I'm currently using are the moog~ that is in pdx, and the plain vcf~.
I might try out muug~ instead. It's working fine with the [switch~]

@Maelstorm said:

They don't. Your computer just gets too small.

I'm ready for a projector. Thanks Maelstorm for the reply.

I guess I'll stick with the moog externals for now. I'm gonna have to go clean up my patches. It's real easy to experimentally add an effect or oscillator here and there when you think of it while playing with these patches. I can imagine the way I want this patch to work but getting there is getting to be quite a challenge.

@billystiltner said:

I guess I should have said vanilla except for my own externals.

That is, in fact, what I thought you meant.

There are some things in PDX that would make things so much nicer like the graphical envelopes, and image, but do all of these work well on all operating systems?

Well, no, but that's not exactly across the board. There's a lot in Pd-extended. The majority of it is cross-platform. If it's an issue, you can always just test on other OS's. Besides, I don't really know of much software that is exactly the same across all platforms.

Is anyone aware that midiout on windows is broken? song position pointer and continue do not work. This was tested not with the latest version of PD but an earlier version. start and stop do work though as well as midi clock.

Case in point. [midiout] is vanilla.