Ray Tracing into a Sparse Voxel Octree

And just when you thought you were through with tracing things all over the place… John Carmack strikes back with a mortal blow with something about ray tracing into a sparse voxel octree!!

The article doesn’t really say much (nothing actually) about the algorithm, and this is where the fun/fuss starts! I can’t wait to see all the amazing/crazy ideas people from all over world will come up with, about what John is actually talking about. Plots over plots will emerge.. flames..

One part that is so standard is the fact that John said he tried this algorithms some years back, even in software rendering… If you look back at when he first mentioned this thoughts about megatexturing you will probably be surprised. It was before Quake 3 was released! I mean this guy is like having a pool of ideas figured out and just sit and wait for the right time to implement.

So let there be fully virtualized geometry.. … done!

carmack,octree,geomentry,virtual,tech 6,voxel

  • remy

    Speaking of virtualization, could we learn how you did your megatexturing implementation please ? The video looked real nice, and the fact that it works over arbitrary geometry is really cool ! Is it available in the sylphis svn ? I’ve been dying to understand how you did it since last august 🙂

    • First, thank you for sharing such indirceble information. : )Second, what do you do for work I’m looking for someone with great graphical expertise to learn from and there are opportunities at various companies in sf that can be arranged. This of course depends on what you want, as I’m sure you have no shortage of opportunity.Third, I am trying to be useful despite my programming ignorance, and have been considering an extension of this that can handle lossy compression. The simplest explanation is to switch to a different coordinate basis (the one used in SH and self shadowing bump maps) and then extend into the space of spherical harmonics rather than Cartesian subdivision. I am supposing error in least significant bits would cause less deviation. It’s the potential, rather than the explicit. This is of course for data that matches (rotations, color space, etc )Is this a plausible approach? I would love to chat about such things if nothing else, it be a great educational thing for me, and I’d be honored too. : ) good work man

  • I found a nice way to implement the megatexturing system, but the description wouldn’t fit in this comment! 😀

    Unfortunately I didn’t finish the system because of “real-work”.. I just didn’t have the time to actually look at the code since August! So the semi-complete code is still in my personal repository..

    I will try to write some post about it when I get the time..

  • remy

    sweet, i’ll be waiting eagerly !

  • there is a nice siggraph08 article by one of the works of id software “jon” so-and-so (not john but jon). should be google-able.

  • iphone5 s c保護フィルム