Not logged in, Join Here! or Log In Below:  
 
News Articles Search    
 


Submitted by Nicholas Chapman, posted on December 13, 2004




Image Description, by Nicholas Chapman



These are some renders from my Monte-carlo ray tracer. The top image took 886 secs to render and uses 3200 primary rays per pixel. The bottom image took 883 secs to render and uses 1600 primary rays per pixel.

In the top image you can see sphere and plane geometry, depth of field, perfect specular reflection, diffuse reflection, and refraction with internal transmission. Specular reflectivity is modulated by a Fresnel term in the red and green spheres.

Monte-carlo ray tracing is a very elegant technique which can handle all GI phenomena, such as specular reflections, caustics etc.. I use russian-roulette ray termination for unbiased estimation of the result of an infinite number of bounces :)

In the bottom image you can see the Stanford Happy Buddha (see http://graphics.stanford.edu/data/3Dscanrep/) rendered with image-based lighting. The model itself is 1,087,716 triangles. Ray/mesh intersection is accelerated with a kd-tree, something that took a good week of coding and tweaking to get working decently :) The env-map used for the image-based lighting is the beach light probe from http://www.debevec.org/Probes/ The env-map is a HDR image, meaning it contains a larger range of intensities than normal 24 bit colour images.

A little bit more info and some more pics are available on my page http://homepages.paradise.net.nz/nickamy/

Nick Chapman aka Ono-Sendai


[prev]
Image of the Day Gallery
www.flipcode.com

[next]

 
Message Center / Reader Comments: ( To Participate in the Discussion, Join the Community )
 
Archive Notice: This thread is old and no longer active. It is here for reference purposes. This thread was created on an older version of the flipcode forums, before the site closed in 2005. Please keep that in mind as you view this thread, as many of the topics and opinions may be outdated.
 
NCCAChris

December 13, 2004, 02:29 PM

Nice piccys! Your implementation is using path-tracing right?
Are you going to implement photon mapping? This will help with the noise and be a heck of a lot faster too :) You are gonna need a hell of a lot more samples so remove the noise otherwise :(

Also - what sort of sampling method are u using? I can recommend hammersley with random offset! It helped loads over stratified sampling!

Good job though!

 
El Pinto Grande

December 14, 2004, 12:56 AM

I think i recognize those pretty pics.

Kudos!

 
Bramz

December 14, 2004, 02:16 AM

aaah, the good old happy buddha :)

very nice!

must ... continue ... my ... own ... raytracer

Bramz

 
Ono-Sendai

December 14, 2004, 04:13 AM

Yeah, it uses path tracing currently, though I'm working on something that's more-or-less photon mapping.

Right now sampling is generally done with plain old random sampling. The only stratified sampling used is for sampling area lights.
I had a read about Hammersley sampling and it looks promising. I'm just not sure how to apply it :) Time to do some more research...

 
This thread contains 4 messages.
 
 
Hosting by Solid Eight Studios, maker of PhotoTangler Collage Maker.