I recently created a hand-drawn VR film in Blender. Because Grease Pencil did not support panoramic renders, I built a camera rig of 6, 90° FOV cameras and batch rendered them. Because of my limited knowledge at the time, I had to use After Effects to import the image sequences, lay them out as cubemaps, and convert them to equirectangular images (scripting was helpful for this process).
While this worked fine, I was hoping to find a Blender-only solution with the compositor. I’ve tried some things with each camera in its own scene, then a composite scene with only the compositing node tree that pulls these in and places them into the cubemap layout for rendering. My trouble is that the individual renders are 1:1 (2160x2160) while the cubemap is 4:3 (8640x6480).
When I try to translate the individual scene outputs, I’m running into this issue where they fill the full render size. Plus, I have trouble getting the Alpha Over node to work like I think it should. I’m all confused at this point. Hopefully someone can help me figure this out (I’ve done a ton of googling at this point, but with no luck)!
…
). So i chose only 256 for the square images (and so 1024x768 for the composite render…). Also i “colored” the images for presentational purposes only…
