Hi, My understanding of glossy lighting vs. increased shininess/brightness: Glossy lighting, if available on your system, has the advantage of making surfaces of all kinds look smoother and shinier even at fairly low vertex density (and ribbons, spheres, sticks at fairly low subdivision quality). However, glossy lighting is not available for some proportion of users. At least in my experience, and as shown in the image comparisons here,
you can get a just-as-nice appearance by increasing the shininess and brightness, but at the cost of also needing to increase the vertex density and subdivision quality. That increases the triangles in the scene and thus computational demands. The publication presets increase molecular surface vertex density, subdivision quality, shininess and brightness -- but they do not adjust vertex density of nonmolecular surfaces. For those other surfaces, how to smooth or if they even can be smoothed depends on how they were created (e.g. volume surfaces could be subdivided and smoothed using options in the Volume Viewer tool) and you would have to do that separately. At least for a volume surface, that is sort of like massaging the data and probably wouldn't be good to have a preset do automatically. So another advantage of the glossy lighting is that it makes all the surfaces look smoother without changing the actual triangulation. The image comparisons page mentioned above doesn't include a comparison of with and without glossy lighting at low vertex density/ subdivision quality, but users with that feature available can try it for themselves. Elaine On Jun 24, 2009, at 3:05 PM, Thomas Goddard wrote:
Hi Elaine,
I agree with your advice. I seldom use raytracing to produce publication images. I do always use glossy lighting. Glossy lighting is not enabled by the publication presets as far as I know. It has to be turned on separately in the Lighting panel (Tools / Viewing Controls / Lighting).
Tom
Elaine Meng wrote:
Dear Fabian, For publication images, to some extent different people will prefer different things. I will describe what I think is important, but keep in mind others may have artistic differences! The User's Guide includes a more comprehensive "image tips" page, also available by clicking the Tips button on the image-saving dialog: <http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/print.html#tips> It seems like many people think POV-Ray is always the fancier/ better option, whereas the Chimera rendering without raytracing only has the advantage of being faster. I disagree. For my own presentation/publication images, I always use the Chimera rendering as I can get much better results that way. This may be due in part to my lack of expertise with POV-Ray, but it is also because there are options only available with the Chimera rendering, and because the shadows from raytracing tend to add to the complexity of an image and make it harder to understand. Of course, the faster turnaround and somewhat more WYSIWYG nature of the Chimera rendering also helps in making nicer images. Most of the Chimera images in the gallery and all currently in the feature highlights page were made directly in Chimera, without raytracing. <http://www.cgl.ucsf.edu/chimera/ImageGallery/> <http://www.cgl.ucsf.edu/chimera/features.html> For images primarily containing opaque molecular surfaces, I would use Chimera (non-raytraced) rendering with settings: white background, increase molecular surface vertex density to 10, turn off depth cueing, turn on sihouette edges, and either use glossy lighting, or if that is not available on your computer, increase the shininess and brightness parameters. ** If you simply use the publication preset #1 or #2 (see Preset menu) it will do all of the above for you! ** Example image from using publication preset #1 is attached at the bottom of this message. Just now, I also made a page with more images showing the settings being changed individually:
However, let's say you have decided to use raytracing because you want shadows. My suggestions for raytracing surfaces would be: (a) increase molecular surface vertex density to make the surface smoother (b) if white background, make the surface some other color (silhouette edges would better demarcate the boundary, but they are not available with raytracing) (c) for faster rendering increase the POV-Ray Option "antialias threshold" from the default of 0.3 to at least 0.5, but 1.0 or even higher may still look as good and be much faster (d) if shadows are too dark, try decreasing the "key-to-fill" ratio in Lighting. Your shadows look much darker than what I got when raytracing today with the default ratio of 2.0. The default used to be higher, but that was a long time ago (changed before production release 1.2540 July 2008). (e) if shadows are in the wrong place, try moving the "key" light position in Lighting The latter two as well as quick shadow location previewing are mentioned in the raytracing page: <http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/raytracing.html> I hope this helps, Elaine