Hello,
I would like the “render pass” that is separated into alpha channel and specular, reflection, transparency, shadow.
Is there a function to control the render layer and pass ?
I’m not sure what you’re asking for, can you be more specific?
Have you tried looking at the scene texture node in the material editor?
Render Layer
For example, we will render characters and backgrounds in separate layers.
In this case,
Character layer:
there will be reflections from the ground and walls upon the surface of the character.
If there is a wall between the character and the light source, a shadow will fall upon the character.
Backgroud layer:
The shadow of the character will fall upon the ground and walls of the background, and in areas that cast reflection, such as glass, you will see reflections of the character.
In this link, you will see divided into individual layers, the foreground airplane and its propellers, the background airplane firing its bullets, and the background.
Render Pass
This function creates output through the extraction of several elements, such as reflection and refraction.
http://core-cg.com/documentation/refpasses.html
We are considering the use of a game engine solution for rendering cinematic VFX(visual effects) and computer animation.
In order to combine with our standard workflows, ultimately it will be necessary for the compositing staff to make fine-tuning adjustments using tools such as Nuke and After Effects.
Therefore, we would like the game engine to output layer and pass separately.
Render layers like you describe do not exist in most 3d game engines, and I suspect there’s no real reason for such a feature to exist either - it would also be pretty costly to try and perform in real time. A large number of game engines also do not have a renderer with multiple accessible passes, though UE4 does thanks to it’s deferred nature - you can see visualisations for each of these passes from within the editor.
Though the feature does not exist to to render out each pass separately, I suspect it would be fairly trivial to implement. If you wanted to implement render layers, I suspect you would need to do a fairly significant amount of work altering the way the renderer works to support your needs - there is unlikely to be a simple solution.
Ultimately though, I question the point of your objectives - you seem to want to use a real time rendering platform and bend it to suit a workflow orientated around offline rendering processes like manually / externally processing and compositing rendered layers. What advantage do you intend to get out of using a game engine like you propose?
I am optimistic about the advantages that you told me about of UE4 compared to other engines.
By using a game engine, we would like to shorten the production time for cinematics and use the time saved to produce more.
Currently existing pipelines using maya and renderman require a lot of time and work.
We do, however, require 2D processes such as composite and edit for the final post processing.
I absolutely support Satoy11’s thread. Using a realtime renderer like UE4 in film production could absolutely accelerate the whole process, as in integrating 3D and live action particularly for instance, but we still need those separate render passes to work in a composite application such as Nuke or After Effects. I cannt picture it’s that hard to get if the math for depth, shadows, specularity, etc. is already being calculated by the engine anyways. There are still other obstacles for UE4 to become truly usable for film production workflow, but getting those render passes would be a good step ahead. I read about some people building a “green screen” in UE4 to then remove in postproduction… Gotta be a better way.
I am sorry for digging such a old thread but this feature would be really useful. Some people are stating that this feature does not exist in most of the game engines and it would be costly but I cannot agree with that. Unity3D and has this feature and it comes with no additional rendering costs. In fact, you can get performance gains if you decrease amount of layers that your camera is rendering. I believe that Cryengine also have this feature.
Possible uses:
- camera that renders FPP hands+weapon
only and composites its render target
over normal game image - camera that renders only some
fractions of the game and is
generating custom rendering buffer
(for example, you want to render
characters only or foliage only, or
just your enemies which are added to
the “enemies” layer)
Will this feature ever materialize?
looks like no, is not even in forward rendering