0

As I put in the title: I struggle to solve that problem. I have a green screen footage that I would like to key and insert in a 3d space. If not on plain, the keying does great but when imported as plain the video frame remains even after keying and the items behind the video footage are not visible (hidden by the video)

Thanks for advice in advance

L G
  • 1
  • 2
  • It is similar but there is not really solution in the description there. The screenshot lacks any further explanation and impossible to reproduce – L G Feb 23 '16 at 11:14
  • The image sequence has correct alpha? How do you import your sequence, via Images as Planes Add-on? Also giving a screenshot as example that doesn't work properly would help. – p2or Feb 23 '16 at 11:55
  • Actually in the meantime I could copy the nodes and it works fine! – L G Feb 23 '16 at 12:20
  • The only bad experience is that if I have a greenscreen with shaded background (a range of green colors to key) in seems very complicated to key out properly. – L G Feb 23 '16 at 12:21
  • was there a solution to use somehow the keying node with planes? – L G Feb 23 '16 at 12:22
  • Chroma keying only works with a uniform color, whether that is a red screen, a blue screen (more common) or a green screen (most common). The color variation needs to be fixed when recording the footage, or you will be plagued by issues the reast of the way through production. – J Sargent Feb 23 '16 at 12:29
  • I am experienced in green screen recording, I am aware of that part. It is just that the video shot in studio is not everywhere the same green, it has variations by the shadow (lamps character not material shadows). I asked if in blender the keying node was possible to use instead of the mentioned in this conversation (color ramp, rgb curve, hue, etc) – L G Feb 23 '16 at 12:33
  • read: http://blender.stackexchange.com/questions/6942/compositing-video-layers-in-3d-space –  Feb 23 '16 at 18:20

1 Answers1

0

Your question has a pretty broad and step intensive answer, but I'm writing this to help point you in the right direction, instead of even trying to give you some sort of magic wand to accomplish what you are asking.

I believe that your answer is to do the keying first, and then get a solid color as your background. You might have to spend a lot of time here as this is an art, and may even require rotoscoping/masking at times, and edge cleanup of your focal subject. If at all possible, you will do most of this work in your footage lighting to reduce the amount of keying problems you would have after the footage is taken.

Once you have a clean video, then your cycles engine can help you accurately isolate your intended subject in 3D space (In pseudo - Real Time).

You have to put in the work somewhere. Cycles is pretty powerful, and can go pretty far, however there is far more power in cleaning up your video first. Then you can use this solid background as a decision factor to mix between video color, and a transparency shader node with a high degree of reliance.

Luckily Blender has the tools to do the cleanup both in the Video Sequence editor for masking, and in the compositor to apply the "VSE" mask and do your chroma/color matte keying.

Hopefully this should point you in the right direction.

Here's a link that I though would help as well: Youtube video showing how both green-screen and masking works

Here's a blend that I put together real quick to show you what I mean by making decisions with color in the cycles engine. The two "colored" images are packed in the file for your study.

enter image description here

Community
  • 1
Rick Riggs
  • 4,623
  • 14
  • 30