PDA

View Full Version : OpenGL 3D Picture Composition Tutorial



vindy
08-10-2012, 11:15 AM
Does anyone know of a website where a good tutorial on 3D picture composition using OpenGL can be found? You know, where you take a 3D mesh data file containing the coordinates of vertices and their edges and construct a picture quality, or near picture quality, image of the object that the mesh represents? I'm not even sure what the technical term is for this, which is why I'm having a hard time finding a tutorial that teaches this using google.

carsten neumann
08-10-2012, 11:57 AM
Isn't that the point of most rendering? ;)
Are you after image quality like the car industry uses for their ads/press releases for example? Extremely photo realistic (still) images are usually generated with some form of ray tracing since those are traditionally better at capturing effects where light is bounced multiple times (e.g. global illumination, reflection, refraction).
To be honest I don't quite understand what you are after, can you provide more details on what you want to do or even what "near picture quality" means for you?

vindy
08-10-2012, 12:28 PM
Hi carsten. Its hard to explain beyond what I have already said. Image quality is what I'm after. What I would like to do for example is take a mesh data file produced from a scan of an object, draw all the faces with realistic looking material (brick, fur, metal, glass, etc), and put the object in a scene of my own creation. I'm sure this would involve things like ray tracing. I've never tried this before and don't really know that much about it, and that includes parts of the process like ray tracing. I've been looking at the PLY image format and think I understand it well enough to make use of it. You can download meshes in this format from http://graphics.stanford.edu/data/3Dscanrep/

There are also images similar to what I'm attempting to do on that webpage. A description of the PLY file format can be found here:

http://local.wasp.uwa.edu.au/~pbourke/Dataformats/ply/

I'd like to do stuff like the still computer generated images put out by Pixar in the past. That's the best I can explain what I want to do. You wouldn't by chance know of any good tutorials on this would you? Thanks for responding to my question.

Alfonse Reinheart
08-10-2012, 01:22 PM
What I would like to do for example is take a mesh data file produced from a scan of an object, draw all the faces with realistic looking material (brick, fur, metal, glass, etc), and put the object in a scene of my own creation.

Then use Photorealistic Renderman (http://en.wikipedia.org/wiki/PhotoRealistic_RenderMan).

OpenGL is a low-level rendering API. It's job is to do exactly what you tell it. It doesn't know what brick, fur, metal, glass, etc are. You must describe these things, using the tools OpenGL provides.

There is no tutorial that will explain every step from mesh file to photorealistic scene. You would need thousands of pages to describe such a thing.

If you want to skip the middleman and go straight from mesh to photorealistic picture, then you don't want OpenGL. You want a professional rendering package.

vindy
08-10-2012, 01:50 PM
Hi Alfonse. I would love to learn how to use Photorealistic Renderman, but for now its way out of my price range. I think maybe I'll try to scale back what I'm attempting to do. I know OpenGL is low-level, but I should still be able to use it to write programs that output a pretty good looking computer generated image of real world objects. Anyhow, thanks for your response.

carsten neumann
08-10-2012, 03:54 PM
A bit cheaper than a renderman license might be to get the book Physically Based Rendering (http://www.amazon.com/Physically-Based-Rendering-Second-Edition/dp/0123750792/ref=sr_1_1?ie=UTF8&qid=1344638827&sr=8-1&keywords=physically+based+rendering). It describes the theory and implementation of a ray tracing system (pictures and code also at www.pbrt.org (http://www.pbrt.org/)).

vindy
08-10-2012, 04:29 PM
Cool! Thanks carsten! I'm gonna buy it.