The holy grail of computer graphics

When it comes to serving up high-quality graphics on a computer, we’ve come a long, long way indeed. From the days of Pong, Atari and the Commodore 64, we now have supercomputing-class boxes in our very home in the form of devices like the PlayStation 3. Even computers, with their graphics card containing fancy components like stream processors that boast of delivering realistic graphics in games and visualization applications, there doesn’t seem to be any slowdown in the pace of innovation in this exciting sector of computing. Computer graphics bigwigs like ATI and Nvidia continue to pursue the honor of being top-dog graphics company by creating more powerful graphics processors, more transistors and ever higher clock speeds.

The story so far
All 3D scenes are composed of little ‘facets’ called polygons—the greater the number of polygons to build a 3D model, the greater the detail. Faster processors enable faster polygon processing, resulting in more detailed scenes that can be rendered quicker. Simple, right? However, this approach is not necessarily one that is extensible, especially given that these days chip fabrication processes are pushing the 55 nanometer range (which is very close to atomic dimensions). So physically, there’s every possibility of running into a stone wall in the near future—it is not possible to fabricate at sub-atomic dimensions. Unless of course scientists make massive leaps in practical quantum physics. But that doesn’t look likely.

The game changer
So along comes a self-taught enthusiast from Australia called Bruce Dell, who proposes a completely different graphics rendering paradigm called Unlimited Detail. He suggests that polygon-processing capabilities should no longer be the baseline of graphics computing prowess. The idea is to treat every point of light on your monitor (yes, the pixels) as the prime element in the 3D scene. In the 3D world, these discrete points are called voxels (many of which make up point clouds). With Unlimited Detail, Dell claims that the rendering engine now acts much like a search engine that basically just decides which points of light to show on the screen, depending on the particular view that needs to be displayed in that 3D scene, the resolution of the screen and other such factors. This approach can drastically cut down the amount of processing power required, as the rendering engine only concerns itself with showing what a user sees. The net result? Hyper-realistic 3D imagery that can be rendered completed in software, with no special hardware support required. I know, it does sound too good to be true. It’s still early days for this technology, but this proof-of-concept demo of Unlimited Detail here provides a compelling argument indeed.

Unlimited Detail - an explanation

There’s another video from the creator of this technology, where he provides a more narrative explanation of how this technology contrasts to the current-day approach.

If this technology does take off, I can’t wait to see God of War 4 in Unlimited Detail. Wow.