koobas.hobune.stream
Culling Mask with Depth Occlusion
So, I've got a system set up. I want to apply a post process effect only to specific objects in the scene. The way I have achieved this is by setting up a second camera, using the Culling Mask setting to render only those objects I want to have this effect. I'm then applying this rendered texture as a post-process overlay on the main camera.
This works, but it has an issue - the Culling Masks don't account for the depth of other objects in the scene, which means that this post process effect is applied even when the objects are occluded by the scene. Obviously for what I'm trying to do, this is unwanted, but I can't seem to work out how to use the render texture from a different camera when rendering this scene, or at the very least how to apply the occlusion from the other camera.
I've tried many different methods, such as using stencil buffers, but I can't seem to get anything to work.
Is there a way I can do this so the effect will be applied only to the visible objects in the scene, rather than all of them?
Okay, so I've done some working around this to make this work.
The first thing I did was use a third camera. This camera has the same settings as my second camera, but with the culling mask reversed - that is, it's set to render everything BUT the objects I want.
Secondly, I applied a replacement shader on my second camera, which would render everything blank white, with no shading or lighting.
Next, I set the third camera to render to the render texture, and render the scene to it (NOTE: this step I would LIKE to be able to avoid if possible, because it does mean rendering the scene twice.)
Then, I clear the render texture using GL.Clear, clearing the colour buffer to black, and leaving the depth buffer intact.
Finally, I render the second camera with its replacement shader. These objects are now writing white pixels onto a solid black colour buffer, but the depth buffer still contains the necessary information.
Now, I have a render texture that contains white pixels where I want the post-process effect to occur, and black pixels elsewhere, which I can use to mask the effect.
This works flawlessly, but it DOES require rendering the entire scene twice. I'm looking into ways I could potentially use the result from the main camera to copy to the second one.
Did you make any progress on this? I'm trying to do the same thing (but honestly I just need the render texture).
