Well lets get this started

My Personal Recording

Week One

I was planning to just do a blog talking about what I was doing each week but then I thought that it would not be that helpful to myself. This blog will not only be a recording of what I have done but also what I have learned and understood.

GPU (Graphics Processor Unit)

The GPU is a data processor, it works very quickly. GPUs used to have a fixed pipeline for handling 

 which is set and cannot be reprogrammed. Nowadays GPUs can do general parallel computation.

Vertex

Each Vertex can have a variety of data stored in or associated to it. This includes: positions, colors, normals and UVs (texture co-ordinates) 

Immediate Mode

Immediate mode is when you use calls such as glRotatef or glTranslatef. These change the data stored instantly. This method is slow though and become depreciated in OpenGL (especially the mobile version OpenGL ES). It is better for your program to use Vertex buffers and arrays in order to store and work with your geometry and glMultMatrixf or shaders for transformations.

To move objects in immediate mode you need to use a transformation matrix. The specific matrix we use is called the model view matrix. We call this through glMatrixMode(GL_MODELVIEW). To project the image we need to use another matrix which is the projection. The projection matrix is essentially where our camera would be in a scene and is called through glMatrixMode(GL_PROJECTION).

Pixels

Pixels make up images and pixels are made up of channels. These channels are RGB and Alpha Luminance. Individual channels are data which generally comes in bytes between 0 and 255. For example the RGB channel values for a bright red could be 255,0,0.

Textures

Textures are images and are generally used just store color. This color is also stored as data. Textures can store other data as well such as maps (see below). Textures are 2D.

UV Mapping

While textures are 2D, models are 3D. We need mapping in order to be able to put our texture on our model. UV points are specified on a texture, one for each vertex on a model, and are linked to this vertex. This means that each UV has an x,y,z position in th digital space. To look up the color we need to sample it which is essentially grabbing the data from a texture that you can use.

Normal Maps

Going back to what we were talking about earlier, textures can store more than just colors. A normal map is a texture that stores normal vectors per pixel. The rbg = x,y,z. This does not change the physical geometry but looks up the normal per pixel and interpolates each pixel giving a complicated appearance to a basic shape by changing how lighting and shading is calculated on a model.

Displacement Maps

Displacement maps are a greyscale image where each pixel stores a value that represents an offset for a vertex or pixel that will move it and change the geometry itself. We use displacement maps in order to overlay a high polygon model over a low one so that we can save on our polygon count ta we have to draw. We can take a 300 polygon model and make it look more like an 800 polygon model by overlaying a more complicated one. Displacement Maps can store vectors instead of just offsets but it generally does not.

Summary

This term is shaping up to be exciting, especially since we are getting into an area that I really want to learn about. As I said I guess I will be using this blog to hold information for myself so that I can go back and study from it. Not sure if I will have to much information. I want to learn shaders so that they can help make up for my lack or artistic skill when it comes to modeling.

Bishop Osborne