First, let's start with the output in QC, because that's how you do
it in QC!
Take whatever composition you have in mind, and put it into a "Render
in Image" (RiI) patch. Hook up that RiI patch to a billboard, so you
can see what you are doing in QC. (I tend to set the billboard width
to 2, so that the image fills up the entire width of the viewer.)
Publish the Image output of the RiI patch, so that your code can get
to it ... see below.
Inside your RiI, whatever image you want to be processed, publish
that input. Notice when you get to the Root, that Image is now an
input to RiI. Publish that input, so that you can get to that input
from your code.
So now your qtz file should have one published input, and one
published output.
Now for the code ....
In my case, I want to do all my processing off-screen, so in my code,
I create an offscreen NSWindow initWithContentRect: ... with some
negative coordinates.
Then, I create a QCView, sized to the desired output image size. I
set the content view of my offscreen window to be my QCView.
Next, you'd hook up your image object to the view, something like this:
