Imagine a globe, with latitude and longitude lines. Each place where the lines interconnect you get a vertice (also known as an angle.) You can deform the globe by grabbing a vertice and pulling it out, down, any direction you want. However, each vertice you add to a model makes it both harder to deal with AND takes longer to compute a full picture.
The water in this animation is just a solid plane. Basically a piece of paper with only four verticies, one at each corner. But then you paint a series of light and dark shadows on the paper. In a 3d Program, the shadows can become anything. they can represent different levels of reflection, or light absorption, color, or in this case, height. Then you cause the shadows to move in one direction or another. The painted image adds about one percent of computation time compared to what it would if you doubled the number of vertices to eight.
In addition I have found ways to tell the 3d program that repetative objects, like carvings on the temple, are just duplicates of one object. The end result of all these changes is that it now takes 6 minutes to compute one frame of animation. It used to take nearly 4 hours. Here is one I did a couple months ago.
