As a total newbie to Unreal and game engines in general, Ive been picking up lots of information from these forums, particularly on how to render cool screenshots and Matinee movies.
I recently experimented with the ‘High resolution screenshot’ function within the viewport dropdown menu. I then ticked the box for ‘Include Object Visualization targets’ and then rendered using the HighResShot 1920x1080 command. I used the Unreal Matinee fight as a test.
The outcome was a nice main image, but then I also got 12 other images with it. My guess is that these 12 images on top of one another make up the main image…
Question is really, what are these images for? and how do you put them all together again if that is the case?
Those 12 images are a breakdown of the shader maps applied. You don’t need to do anything with them. They are for reference. It’s to help you pinpoint how your shaders are working and to see if they are working how you want them to.
Unreal Engine 4 uses deferred rendering, this is a technique where information of the scene gets rendered into multiple render targets (the GBuffer) which then later can be used again for post processing (which also includes lighting).
Most of what you see in those different images are visualizations of the GBuffer and usually are for debugging and development purposes only. But the scene depth for instance could be used to reconstruct the 3d information and thus be used for 3d images/movies and stuff
Better to think of them as data that exists for each pixel. The images are just representations of that data that make sense to humans.
Some descriptions of the ones that might be relevant to you in the near future as you learn the engine (Especially making Materials):
Diffuse: What color should each pixel be if there is no lighting/shadows/reflections/etc… applied?
Roughness: How well does light reflect for each pixel? Black is perfectly reflective, white is not reflective at all. The Epic tutorials on Materials found on YouTube describe glass as being black (well, slightly lighter) and chalk as white (again, actually it’s slightly lighter).
Normal: What direction (in 3D) is each pixel facing? When you use normal maps in materials, you can fake details in 3D for surfaces that would otherwise be perfectly flat. Helpful for things like brick walls that are flat as far as the game cares, but should look 3D to the player.
Remember that all the maps in color are three values per pixel, usually from 0 to 1. And all the black and white maps are one value per pixel, again usually from 0 to 1. But since this information exists for every pixel, why not actually display that information as a picture so we can better visualize it?
