Raytracing with POV-Ray

The Persistence of Vision Raytracer (POV-Ray) is included with Chimera. POV-Ray images can be saved:

with Save Image
with the copy command
when making movies

Raytraced images can only be saved in the PNG format. For each raytraced image, Chimera first generates two POV-Ray input files:

scene description (*.pov file)
raytracing options (*.ini file)

The raytracing calculation is then run as a background job that can be monitored or canceled using the Task Panel. Several raytracing parameters can be set in the POV-Ray Options preferences.

Lights and Shadows

Chimera includes two light sources, key (dominant) and fill (secondary). Their directions and intensities can be adjusted in the Lighting tool and are propogated to POV-Ray for raytracing. By default, only the key light produces shiny highlights and shadows. Decreasing the key-to-fill ratio will decrease the contrast between shadows and unshadowed regions.

Interactive shadowing in Chimera (the show shadows setting in Effects) can be used to preview where shadows will fall.

Balancing Time Requirements and Results

Sometimes a considerable speedup in raytracing can be obtained with little change in output appearance by adjusting settings in the POV-Ray Options preferences. This is particularly true for movies, which are compressed to some extent during encoding. Changes to consider include:

decreasing the quality to 4 or 5 [quality details at the POV-Ray site]
increasing the antialias threshold to 1.5 or even higher (maximum 3.0) [antialias details at the POV-Ray site]
if objects are not intentionally sliced, making sure that global clipping is turned off; to adjust depth cueing, use the start and end parameters in Effects rather than moving the global clipping planes, which would turn clipping on

There is no single answer as to which values will optimally balance time requirements and results, as it will depend on the contents of the scene, the available computational power, and the patience of the user. Running a few tests is recommended before committing to raytracing a large set of images with a particular set of parameters.

Raytracing jobs can be monitored and canceled using the Task Panel.

Limitations

Using transparency or clipping may increase rendering time significantly.

A raytraced image from POV-Ray may differ from the view in Chimera in several ways. Some differences are desirable, such as the presence of shadows. Others reflect current limitations of the POV-Ray format and/or its implementation in Chimera:

colors may differ slightly
shading along ribbons may be stairstepped; using the supersmooth style instead of rounded reduces this effect
light source directions but not colors are translated from Chimera
although object shininess and brightness levels (but not specular color) are translated from Chimera, objects tend to appear shinier in the raytraced image (perhaps a desired effect)
if depth cueing is used, the background of the raytraced image:
- will not be transparent even if the transparent background option is turned on
- will match the depth-cueing color even if the background is a different color in Chimera
the following are omitted or otherwise not handled: silhouettes, selection highlighting, stereo, Volume Viewer transparent solid displays, and normals on dot and mesh surfaces
max_trace_level is set to 10 to improve multi-level transparency, but this also increases time and memory requirements relative to lower levels [details at the POV-Ray site]

UCSF Computer Graphics Laboratory / August 2009