Don’t really know how these should be implemented ?
As options on some 3D cameras Objects ?
How GPU heavy would this be for non-offline (real-time) compositions ?
Opened for voting.
1 Like
Can be achieved in GLSL using available depth map.
Maybe this would be a good way to implement to make it as modular as possible. (As opposed to adding a setting to a camera for example)
I saw an example where a developer made lens style depth blurring as opposed to smooth Gaussian blurring, very nice!
@alexmitchellmus, we’re considering various ways of implementing depth of field.
(a) Node that inputs the Depth Image output from Render Scene to Image. As you said, this has the advantage of being modular. A possible disadvantage is that you couldn’t use it with Render Scene to Window; you’d have to refactor your composition to use Render Scene to Image.
(b) Setting on camera nodes, with the option to disable it.
(c) Node that, like the camera and lighting nodes, you connect to Render Scene to Image/Window to enable depth of field.