Compositing Research

Compositing is basically the process of embedding CG with real life footage. It is the manipulation of mathematical formulas such as color etc. There are many things to consider in compositing such as the camera that you use. The settings and specs of the camera can drastically change the result of the composite. For example, most cameras use CMOS chips which distort the pixels of the shot. This is due to the fact that the camera 'reads' the image from the top to the bottom bringing about a sense of 'lag.'

Another important thing to note is that because CRT monitors were in use a long time ago. This made iamges darker than what they should appear. So all images are embedded with a +2.2 gamma difference and hence SRGB color was created.

When compositing using programs such as nuke we want to work in a linear light space but export our final result in SRGB light space. Another thing to consider before exporting is the camera distortion. All cameras create a distorted view of the environment and this is not very useful when processing pixel information because the computer requires a 'perfect world' work with. By undistorting the image we remove the effect the lens has on the image and we can proceed with embedding CG into our footage. At the end of compositing we can then re-distort the footage to bring about the desired result.

When bringing in CG to nuke we render out multiple passes such as ambient occlusion, reflection, lights, shadows, diffuse etc etc. These are all the layers that are to be manipulated into tricking the human eye into thinking that the foreign object belongs in the environment being shown. In order to place and animate the CG we motion track the footage. This is a very tedious and time consuming process because you must selectively choose areas that you wish to track. You can mask out areas that are undesirable such as shadows and moving objects. This is done so that you can calculate the 3d environment as accurately as possible. When the scene is calculated into 3d, you import it into your 3d application of choice where you can begin setting up your CG and rendering.

Once again the process is all about tricking the mind into believing the shot that is presented to them.