Silly OpenGL Shader Program I Wrote

[Incomplete Dude]

It's based off the idea of rendering using raymarching on distance functions, except I bastardized it to render metaballs. Take a look if you like, the magic is in dffs.txt. I should probably rewrite it to use the newton-rapheson method for root finding, but oh well.

Disclaimer: No warranty implied. If it destroys your computer, tough luck :p

[kevintje]

can`t open it...

it sais the application configuration is not correct, and that reinstalling the program MIGHT solve the problem...

are you sure it`s winXP compatible?

[mm3]

O___O It BSOD'd one of my laptops. Not gonna try it on another one! x_x

[Incomplete Dude]

Originally Posted by kevintje

can`t open it...

it sais the application configuration is not correct, and that reinstalling the program MIGHT solve the problem...

are you sure it`s winXP compatible?

Of course it is, I didn't do anything special to it. I really don't know why it doesn't work.

I code it's based on comes from this program http://www.pouet.net/prod.php?which=49796 So if that works, then I'm at a loss as to why this one doesn't.

Oh, it also requires PS3.0, so if you have a GPU older than Geforce 6000 or Radeon x1000 series, it won't work.

[kevintje]

I tried it again on my own computer at home, same problem, can't open it cuz the configuration the program is not correct, reinstalling the application might solve the problem.

I guess you need additional software/libraries to open it :s

[Incomplete Dude]

Ok, I checked exactly what dlls the program loads, and I've changed it to statically link the ones people might not have. If someone could be so kind as to try it.

Also, because of this, it's now too big to attach here, try this link:
http://www.mediafire.com/?wgmh3j2mnjc

[chevre]

post some screenshots? We Linux users can't try it

[Incomplete Dude]

ok, well, the PrtScn button simply refuses to work while OpenGL is running. I think it's a Vista thing. Either way, I used the corona image library to manually take a screenshot, and here it is.

Some explaination. The background is simply abs(viewDirectionNormal), but on the object green indicate a greater number of iterations used in evaluating the surface, while red is standard phong lighting. You might also notice some slight banding on the object, but that's simply because my algorithm is a major hack lacking in accuracy and precision.

EDIT: Change the photo host to imageshack because the forum bastardizes the image too much.
ImageShack - Hosting :: imagenv7.png

[Incomplete Dude]

Ok, I did some improvements. I drastically improved the performance and accuracy by using newton's method. There is no longer a penalty for finding edges, computation expense is pretty uniform. The previous version was doing in 64 steps what I now do in 16. I've also added shadowing, ambient occlusion, more complex geometry and screenshot capability. Attached are some screenshots, again green indicating the number of iterations required to find the surface.

ImageShack - Hosting :: image2hk4.png
ImageShack - Hosting :: image9nk2.png

[kevintje]

is that flubber? O.o

anyway, I still can't open the compiled version (it opens, but gives a white screen and stops responding), also, you included main.c, but to compile it myself, it needs to have GL/glew.h too, wich is not included

seems nice tho'