I had an idea that I could use the user_defined_camera in Megapov to give me a
better focal blur than I was getting with my existing Camera35mm macros in
"stochastic rendering" mode, so I started hacking together some functions that
worked with the feature.

Long story short, I suck at maths.

But... I'm almost there. The basic premise (use the user_defined_camera to vary
every pixels focal blur offset) seems to work very well. Compared to the POV
focal blur with very high settings, I tend to prefer the user_defined_camera
version in most parts of the image, and I know that more passes would improve
any remaining artifacts.

I'm still struggling with only being able to use functions (which I've just
started to learn more about) and my maths is, as I said, pretty poor. At the
moment this code only works along the Z axis with the focal plane at z=0, but
I'm on the home stretch now, and I should be able to do the final work to set an
arbitrary "look_at" and "location", as long as I can get my head around the
relevant transform functions.

So, umm, is there any chance we could get the "user_defined_camera" into POV
3.7? It always struck me as the right way to add arbitrary new camera types to
POV, as opposed to having to change the actual code, and this experiment, to me
at least, certainly confirms this premise.

Cheers,
Edouard.

Post a reply to this message

Attachments:
Download 'multipass_plus_user_defined_camera.jpg' (114 KB)

Preview of image 'multipass_plus_user_defined_camera.jpg'