Mortze wrote:Putting it simple, the rendering process ends with the graphic card processor defining what RGB (Red-Green-Blue) colour value goes for each pixel in the picture. It does that with complex calculations. The more pixels in a picture, the more calculations the GPU has to make. More time to render.
It's a little bit more complicated than this. If you render the exact same scene in two resolutions, the one with more pixels will take longer. But a lot of things have a bigger impact on the render time than the number of pixels. Scene complexity is much bigger. Render a scene with one model in front of a white background, and it'll go faster than a scene with that model in a 3d rendered bookstore with racks of books behind her. That'll be true even if you render the simple image at a higher resolution than the complex one. Similarly, lights can have a big impact on render time.
Note that a flat jpg background doesn't really add to the render time. So if you want to render something with a really complex background, sometimes it's better to do it in multiple steps. For instance, say you want to render an image of a lady dancing at a nightclub with 10 other people behind her. You could put all 11 people into the scene, tell it to render, and then come back in a few hours. Or you could render the scene with just one or two of the background people and save it as a PNG. Then remove those two people from the scene, make the PNG be the background, and add two more people to the scene and render it. Now use THAT PNG as the background, remove the two, add a few more, etc. In this way you can build up the complexity of the scene with more quicker renders. It won't be perfect, because the shadows from one couple won't interact with the later people, but for some situations, it might save you a lot of time. Also, if you do a couple backgrounds this way, you might be able to get away with using them for a larger number of foreground images. Imagine 3 backgrounds without the main character in it with people dancing in different poses. Then make 10 images with the main character that cycle through those 3 backgrounds. It's not as good as fully posing everything every time, but it's not bad either. Most of the Keeley-verse games used stock images that weren't even rendered. Things are
way better with Mortze's fully rendered scenes, but the Keeley-verse games were acceptable.
Tlaero