Use OpenFrameWorks/Quartz Composition image data?

Question

So I have been tinkering with the Skeleton sample project on a Mac and I have a couple of questions. First off, How can we pass image data from say Quartz Composition or a Openframeworks project(C++,Xcode) or any other to the render function of the plugin. For example from Openframeworks I can get a pixels as an array of unsigned chars. Can I supply it to the render function? If so, how? Also I have seen people using Quartz Compositions to make plugins for AE. Can some one help me figuring out how to pass the image data from the composition to the render function in a plugin? Please bare with me, as I am a newbie and I really hope some one could help me.
Thank you!

Answer 1

as for bit depths:
there a 3 different iteration suites. one for each bit depth.
i couldn't find them in the documentation but if you go to the definition of any of them in the code, you'll find the others.
you must use the appropriate suite for the kind of buffer you're using or the function will fail.
how do you tell which one to use?
if you created the pixel world yourself, then you know it's kind.
to find out which depth AE hands you, use PF_GetPixelFormat.
take a look at the "smarty pants" sample project. it shows how to handle that.
actually the shifter sample implements that as well, so forget what i just said. (yes you were right)
as for your x y question, i'm not sure what you're asking.
i managed to understand that question in 2 ways.
so here are 2 answers.  :-)
1. if you're asking how to tell the coordinates of the pixel you're currently being handed by the iteration suite,
then the x and y coordinates are given to you by the iteration function.
to be more precise, it gives you the following data:
refcon - a pointer to whatever piece of memory you sent to the iteration function. (use it to transfer data into your iteration function)
x - the x coordinate of the current pixel in the output world.
y - that one's kind of obvious...
in - the current pixel's content from the input world (if the in and out worlds are of the same size, then the x and y coordinates for both worlds are the same)
out - the structure into which you put your result RGBA (you may fill that world with whatever you like before the iteration, thus getting two inputs to use for your calculation)
2. if you're asking how to tell which pixels are being sampled using the sample suite (as used in the shifter sample),
then these are shorts, bit shifted.
these values are compatible with what you get reading data from a point param.
these values are NOT comfortable to use for anything else.
why? relics of what was most efficient CPU-wise 12 years ago.
you can sample a pixel buffer in 3 ways:
"nearest neighbor" for fast performance,
"subpixel" for accurate results,
and "area sample" for... when you need... an area. (duh)
for more info, investigate the "PF_SampPB" structure.
in any case,
you don't have to get the input using the sample suite.
you can always use the input pixel as it is, or get the pixel from the quartz thingie you want to use.
i tend to think that the first answer was what you we're looking for , but what the heck.