Put a signal multiplier on every signal path, and figure out the switching logic to just turn on the ones you want.

More info

Just changing window focus fixes the problem caused by moving the window. For example, if I switch back to the main pd window then back to the patch, the plugin does it's job. I suppose it just isn't being called on a window reposition.

Also, I grabbed the 0.51 source and compiled it - and it behaved the same way as 0.49.

I have reverted back (because LADSPA plugins were not working in the 0.51 version I created. Is there something special I have to compile in, or change, to be able to use LADSPA plugins?

More info. On a different (linux mint) computer, I have Pure-data Pd 0.48.1 installed, and the gui plugin behaves properly. Something changed in 0.49.0?

Can I get 0.48.1 for raspberry pi from anywhere?

I am trying a GUI plugin called editmode_look-plugin (downloaded yesterday from the puredata site. It works, but when I move the loaded patch window on screen, it reverts to the standard look until I go into edit mode and back out. It also occasionally starts with the standard look when the patch file is opened, again, going to edit mode and back fixes it.

This is on Raspbian, Pure-data version Pd 0.49.0

@seb-harmonik.ar said:

@pure_raspberry the frequency setting only goes up to sr/2pi, so in your case ~7639.44
see here for a full explanation about the frequency-warping and patch https://forum.pdpatchrepo.info/topic/9723/lop-object-argument

OK, I think I am understanding it. Is this just a limit for the setting? If I set a low pass filter to , say, 7kHz, then the response still looks like a normal low pass one pole filter, slowly rolling off , even at frequencies above 7639.44 ?

I am using lop~ and hip~ filters in a real time audio application, at a sample rate of 48kHz.

Something I read in the help of an external library has me worried they may not be working as I expected.

A variant of [lop~] that has a signal-controlled cutoff. Like [lop~], the filter has no effect when the cutoff is above SR/(2pi)

Is this saying the filters are only working to about 5kHz?