Welcome to the visually pleasing world of Anti-Aliased Splatting

This work describes three new results for volume rendering algorithms utilizing splatting. Firstly, an anti-aliasing extension to the basic splatting algorithm is introduced that mitigates the spatial aliasing for high-resolution volumes. Aliasing can be severe for high-resolution volumes, or volumes where a high depth of field leads to converging samples along the perspective axis. Anti-aliasing is achieved by prgressively increasing the size of the splatting kernels, starting from a certain distance from the viewing plane. This has a lowpassing effect for the volume portions that would otherwise be rendered above their Nyquyist limit. Secondly, an analysis of the common approximation errors in the splatting process for perspective viewing is presented. In this context, we give different implementations, distinguished by efficiency and accuracy, for adding the splat contributions to the image plane. We then present new results in controlling the splatting errors and also show their behavior in the framework of our new anti-aliasing technique. Finally, current work in progress on extensions to splatting for temporal aliasing will be demonstrated. Here, we present a simple, but highly effective scheme for adding motion blur to fast moving volumes.

Want to know more about it?


Perspective view into a checker-boarded tunnel, using standard splatting kernels:

Notice the strong aliasing behind the black square.

Perspective view into the same tunnel, but using the new adaptive splatting kernels:

The strong aliasing has been smoothed out, as a function of depth.

Terrain perspectively viewed with standard splatting kernels:

Click here for an aliased flight over this lovely terrain. Notice the crawling ant effect in the regions further away from the viewpoint.

Terrain perspectively viewed with the new adaptive splatting kernels:

Click here for a quicktime movie of the fly-over. Notice that the crawling ant effect has been replaced by a slightly blurred and much more pleasing appearance of the terrain, caused by the lowpassing of those volume regions. Click here for a movie that shows the distance k , illustrated by a black line, from which on the size of the splatting kernel is progressively increased, i.e., lowpassing is performed. The volume regions in front of the black line, towards the observer, are not lowpassed at all.

To Research Homepage