New technology brought to us by Bruce Robert Dell and Greg Douglas allows games to render unlimited high-poly point cloud system models in real time, through software.
That means no more quad core to play all the new games.
Not even a dual core.
You don't even need a high end processor at all.
Or a graphics card, onboard will do fine.
Heck, if Crysis used this technology, anyone would be able to play it on their grandma's laptops.
Here's how it works.
We currently use polygons for our 3d models in games.
This means we have to render lots and lots of sides for every model, like facets on a gem.
The more facets a gem has, the more rounded it is.
Everyone wants a tree to have a round trunk in a game, so the tree needs lots and lots of facets, which takes time and power to render.
Because of this, graphics card companies have been rolling out new models every 6 months, and raking dough in the other side.
If video games used a point data system instead of polygon system, it would allow for photorealistic models with no sides at all, incredible detail.
Instead of a wall with a texture of bricks on it, to create the illusion of a brick wall, points would allow the wall to actually be built from tiny brick models.
It would look incredible.
The problem?
Graphics cards aren't there yet.
Rendering every single point in a scene takes herculean power.
It takes hours to days to render a single picture using points, and because of this points are only used for artwork rather than animated movies or games.
With Unlimited Detail technology, computer software will only display as many points needed to fill your screen, based on its resolution.
For example, my screen is 1240x1024.
That's already math, so we know there are 1,269,760 pixels on my screen.
Now, rather than rendering the billions of points for all the models in a room of your video game, Unlimited Detail will calculate which points you actually need to fill all 1,269,760 pixels on your screen.
This means you'd only need to render a little over a million points per frame, instead of billions.
Thus, time and energy is saved, allowing smooth gameplay with unlimited detail.


3 Responses so far.

  1. O hai chabitsku

    nice article

  2. It has been discussed a lot by graphic artists and game developers and really instead of making things good it makes it harder to make for example: raytracing for shadows and reflections needs geometry to work, since this technology only loads the pixels you need there is a lot of data lost so raytracing is impossible.
    for animation there is the problem with physics, since you don't load geometry anymore there cant be collision calculations so there wont be physics anymore.

    here is a video review that presents that unlimited detail

  3. Like the video says, raytracing is slow anyway, and is becoming obsolete.
    There are other methods of calculating shadows and lighting that work just fine with UDT.
    I'd also like to correct you on the geometry part, points do indeed have geometry, and can be animated/have physics applies.
    All that was covered in the video.

Post a Comment