I’ve seen a lot of Youtubers mentioning in their videos that “This took so-so many hours to render” ,and I don’t understand if something is visible as it is before render why is it so much time-consuming to extract it in that form itself without rendering. I know I might be sounding so dumb but I really need to know how things get rendered be it a video or VFX shots.
Thanks.
In: 2
This depends on the sort of content you’re viewing. In the case of 3D VFX usually artists will be working from a low resolution model with fewer vertices while animating. They’re usually also not working with final textures or lighting. Then whenever they hit render the low res versions of textures and models get switched out with a high res version. Then of course there’s lighting—usually if you want photoreal lighting you’ll use a process called raytracing. As the name implies, this means when you hit render that the computer “traces” the rays of light and calculates the way light would bounce off of the objects in the scene. This is very processor intensive as one can imagine because light bounces a LOT. Then of course there are things like ambient occlusion, depth of field, and subsurface scattering that have to be calculated.
Oftentimes to save time a studio will split these different calculations into passes which then get rendered out separately or evenly distributed across a render farm. Then it will be brought back together in the composition phase.
In the case of video without VFX, usually editors will be working off of low resolution proxy files that have to be swapped out for the higher resolution ones that contain more data (usually in the color space). Color grading and color correction is then applied to the full resolution file and exported.
Note that the time it takes for the computer to do these calculations will balloon the higher resolution you want the final product to be.
Image quality is the problem. Ever played a 3d video game and thought “wow, those shadows on the moving players/characters/monsters look like crap!”
Well, good quality rendering will actually fire rays of light from the light sources to the objects and trace their exact shadows. And it is mindful of multiple light sources, how they are brighter in some areas and darker in some areas, and produce proper illumination at all locations, shadow or otherwise.
And that’s just the tip of the ice berg. If you’re in a room with 2 walls painted different colours… let’s say one is white but the other is a blue colour… the white on the edge of the blue wall looks slightly blue because light is reflecting off the blue wall and that reflection is blue light. A “good render” will do exactly that. Similarly, if the light coming into the room is a shade of red, the wall will tend to look darker because a blue wall doesn’t reflect red light very much.
In short, light bounces around a lot, and realistic lighting must take that into account. And it’s a LOT of work.
These are just some of the subtleties of light that your graphics card doesn’t process when you need 60 frames per second or more. In fact it can take hours to produce a single image that properly accounts for all these things.
(And the above list is by no means complete. Materials that are polished act a bit like mirrors for light, dull edges do scatter light a bit, and so on and so forth. Everything reflects light at least a little bit, or else it would look black)
In short, there’s a ton of math involved. Briefly, to render a scene, you have some mathematical model of what is in the scene, the objects, lighting, light sources, reflectivity, etc. Then, the rendering program calculates what you see at each pixel in the rendered image of the scene, tracing the virtual ray of light backwards to see what it hits and figure our what color that pixel will be. It hit the red ball, but it’s kinda shadowed from the light source, so it’s a dark red color. Then it does that for the next pixel and so on. It’s all math. 🙂
Latest Answers