I have a simple piece of footage of a 3D character in front of a “green screen”. I then have a ‘background’ image of a road. I have added both to the Sequence Editor and used this greenscreen plug so the background appears behind the character -
All this works fine and renders fine. Now I want to add shadows - for the Character at front to give shadows on the road behind in post processing. Does this need to be done with the original character video when it was rendered out first? Remember the objects in this scene are a static background image and a green screened video.
Just thinking off the top of my head. You could use your alpha mask (black/white) of your foreground animation after you pull out the green and map the black outline on a plane above the ground and do some kind of alpha blend of this with the ground. you could also make a ground plane with the mask as an animated texture which is mixed with your ground texture or still image of the ground.
Start with making a black outline animation of your foreground (on white) or the opposite. That will be the shadow. Whether you use white on black or black on white will depend on how you mix the texture on the road or the background (darken vs screen etc) Then use this as an animated texture on the ground. Or if you just want it on the background map the animation to a plane that the foreground character will “stand” on or stand in front of. It should be transparent except for the character’s shadow which will either darken or mix with the plane. Play with the perspective of the mapping in the texture panels or the position of your transparent plane. (depending on which method you chose)
If you post more details about what your setup consists of it would be helpful. There are several ways of doing this as you can tell.
Here you can see the JPG image in the background and then the bot character doing a 360 in one position in the front. The latter is/was a green screened animation [ AVI also ] [ bot with green background ] - The above plug got rid of the green to show the background road. Both the still and orig AVI added to Blender Sequencer.
Make a plane that matches the perspective of your road and give it a material that has Only Shadow enabled on the shaders tab. Now, move the palne to a seperate layer from the rest of the scene and be sure that all layers are visible. Create a new layer and name it Plane or something to that effect. On the second row of buttons in the render layers tab be sure that the layer containing the plane only layer is visible. This will cause this layer to render with only the plane containing shadows from all other objects in the scene. If you click the Single button on the same tab only the shadow plane will render, otherwise the two render layers will be combined in final output. You can render with file out put nodes so that the two render layers (shadow and character are rendered to separate file outputs and even different formats if you wish but will be rendered to those files simultaneously). Setting things up this way will keep you from having to re-render the character while still getting the shadows. Blender’s render layers are one of the coolest things I have ever worked with on a computer. They are extremely advanced and can be just as confusing until you grow accustomed to and familiar with their use. They will however allow you to Bitch slap Blender into doing almost anything you can imaging, things which would otherwise be extremely time consuming or outright impossible. They’re very much like Blender’s scenes but can do things that scenes can not do.
Sounds like RamboBaby has roughly the same idea. Make a plane that only adds a shadow (click on Only Shadow) However I was suggesting using the compositor, but you could also just use render layers.
aren’t render layers awesome? I’m not sure about bitch-slapping Blender though…I feel sometimes like Blender bitch-slaps me!
No, once you really start to understand the compositor you can tell it to do just about anything and it will obey like a desperate little trick who will do anything to please you. OK, I’m just funnin’ ya now. It might be a little piece of leather but it is WELL put together. I don’t know of anything quite like Blender’s compositing system. The decision to make it a combination of node and layer based was just brilliant. It makes it a pleasure to use because it’s so incredibly versatile.
So one creates another image sequence of the character shape in white on a black background, this is the ‘shadow’ sequence - [ call this SEQ2 ] as well as the original image sequence with the green background keyed out [ call this SEQ1 ]. Now you have two sets of image sequences.
In Blender you have three planes, background plane has the Road Image, foreground plane has SEQ1 and a third plane has SEQ2 [ the shadow sequence ].
Have I went off down a side road here?
As an aside - if one loads the ‘green’ image sequence into a plane as a texture in Blender, can one key out the green background in the workspace, clip it out?