[gyc]

Hi,
I'm not sure where to put this, but can anyone point me to a general algorithm that will enable me to plot a set of vertices connected by edges to the screen. I have a set of vertices w/ x, y, z coordinates and 3 vertices each are connected by 3 edges to form a triangle. Is there a general algorithm to transform the resulting 3-D image into pixels that I can plot on the screen?

[00101001]

I've been working on a 3D engine in my spare time, so I might be able to help you. First of all, do you want this projection to be perspective correct, or isometric? Does it depend on the rotation of the camera of polygon? Is it shaded or just the outline?

Lets say that the camera was at (0,0,0) facing straight down the z-axis. To transform the coords of point A from 3D to 2D space, the transformed "screen" x value would be : A.x * k / A.z.

The y value would be A.y * k / A.z

where k is constant depending on your scene.
do this for all the points adn then connect them with lines.

This makes sense if you think about it, you are taking the hor/vert value and dividing it by its distance from the camera. There is some really awesome other stuff you can do in 3D, sorry if this was vague. I have a simple 3D engine that I had written in Java, you can check out some applets (they are kinda old) and see the source here:
http://homepage.mac.com/lorenb/

check out the source that says q3d...

anyway, it is late, and my grammar is failing me...so please either post or email any questions you may have.

[parallax]

If you want _the_ definitive source on 3-D programming, it's Foley's Computer Graphics: Principles and Practices. Although it's old, it's probably the best graphics book out there.

[ktolis]

I would suggest not trying to make your own implementation of a 3d graphics engine. Instead try using OpenGL. I am curerntly working on a physics model that will be using OpenGL as the rendering engine. This way things get very simple and I can focus in what I want to do and not fiddle around with maths for 3d->2d projection.

If you want to know the way things work is that you always have the camera look from -z towards 0,0,0. So you always normalize the 3D coordinates to the "world coordinates" so you can have one algorithm for drawing everything.

ktolis

------------------