Correct Application of Sub-patching - all complicated patches?


#1

Hey all,

Trying to work out the do's/dont's of subpatching. I have a pretty complicated patch going on with multiple effects blocks. I'd like to hide all of the maths, scaling, constant objects and just have subpatches of each affect with the key parameters on the parent. Is this a good thing to do, or does it lead to complications? I've read a few threads of people losing connection to sub-patches, but not really sure how/when this occurs?


#2

Hi,

I can think of it as putting all that stuff inside that would serve one function. E.g. you want to make a synth. A synth can be broken down into several components: midi input, oscillators, envelope generators, lfo's, fx... Each can then be one object dedicated to a specific group of functions. Then you could again break down each of them into their components, e.g. Keyboard input can consist of gate, key track, velocity, pitch, vibrato, ... whatever you want to have inside them. e.g you want to add features to the adsr factory object, like invert or keytracking, curves, you could just build a patch where you simply hide all those components and outside it will just look like the adsr before with added features.
A hardware midi controller with knobs, sliders etc. can be useful too, you won't have to display those inside the patcher or have cables running all over your patch, they can also save a bit of cpu or ram ( or both, if someone could correct me please. Webmidi if you don't have the hardware or an iPad, there are several apps, if you know pure data you can get pd-party and make all your gui there and use that as your midi controller, it's open source.
What I would avoid is approaching the patcher as an interface, depending on knobs and controls inside the patcher. I think it's was not designed for that purpose.
Hiding math operations: your function needs several objects, const or operational or conditional objects. It's actually not all that scary to go inside those and change them to your needs. I have very poor c programming skills but it didn't take long to figure out how to modify these and I could have one object at the end instead of 7. hacking objects 101 is worth a read.
It would also help to post your patch and others can have a look and help you with it. I'm sure there are more ideas out there, this is what I came up with, so it's not really a rule.


#3

Problem with stuffing too much in one sub patch is that they get really hiiiiiiiiiiiiigh and don't fit in screen.
For now knobs can only be alligned in one way, vertically.

If you are just making a simple synth, what Igel suggest, seems fine, separating it into envelope , filter, osc, etc.

But if you are making a "multipatch" with evrything like kick/bass/drums/Pads/etc. in same Axoloti, you really have to be smart about how you use your DSP and memory ressources. Check this post for updating that anf also see what else you can do to save resources for complicated patches.

And actually if you look in to community library you can find many objects where users have actually all ready combined several objects into a single object. Which then removes the need for a subpatch. An example could be an envelope that I made, look for jaffa/env/adsr1, its a combination of several objects, made like that to avoid using a subpatch:

As igel also says, try avoid having to many knobs and stuff like that inside the subpatches, only what you need. Cause even though you dont see them they still use memory.