And Rendering...

Once Alpha was done we pulled an all nighter to get to the rendering phase once again for our final stage! We fixed up our animation and added some more environment elements. As today is a statuary holiday we have spent it in the labs taking up every computer we can to render. Literally every computer we can: we just hit the maximum available Maya licenses for the school. So now we are waiting for the smaller files to stop rendering so that we can begin in new ones right away.

As we don't really know how to render properly we are taking a risk in the way we are doing it. Since our files are way too large and complex to render on their own anymore, we've separated each one into different layers to render individually. For instance, the grass in Scene 1 will not render as a single field, it is broken down into 10 smaller patches. We have looked into different methods to do grass but found that it would be a bigger toll to redo it than to continue on with what we have and hope for the best.

In order to render the different components separately, we couldn't just break them apart. Each of the 10 layers of grass that is rendering needs to also have shadows being cast from all the different environment objects and characters. To do this, every object is present in the scene, but we have set it so that only the ones we need to be shown for the current pass is shown, with the rest being rendered as invisible shadow-casting objects.

To speed up our renders as much as possible, we are skimping a bit on flexibility. When rendering, each frame comes out as a set of "passes", which contain all of the information from a certain aspect of the master render, such as shadows, highlights, or even more complex data like motion vectors and relative-camera depth. The more render passes you produce, the more control you have over the components of your scene, but each render pass added takes up additional time and space per computer. To minimize the time spent rendering, we chose to only render out the passes we think are necessary in each scene, often using only the master and a depth pass to help in compositing them together.

In addition to all of our characters and environment objects, we are also rendering out a layer of coloured fog with each scene. To avoid getting too wrapped up in the details and complexities of volume shaders, particulates, and noise generation, this fog has to be rendered using a completely different renderer (Maya Software as opposed to Mental Ray). To make our fog, we simply created a large cube around the entire environment and attached it to a a Maya volume. After playing around with some settings and colours, we achieved something which gave a fairly realistic distribution. This fog layer will be useful when compositing, because we can use it to give the environment a much greater sense of depth and occlusion.