Povscript+ FAQ


Q: my labels are not rendered correctly with raster3d! Why not?

A: As of povscript+ version 2.02, you need to use the "label3d" program in the Raster3D package, as povscript+ uses object types 10/11/12 to produce (higher quality) labels, as opposed to the old molscript style output. Also, be aware that newer versions of imagemagick (>6.0) don't work with the current version of label3d. To fix it, edit the label3d script and change the following:
find (around line 122):

   convert -draw "matte 0,0 replace" ${tmp}_label3d.png ${tmp}_label3d.miff
change this to:
   convert -fill none -draw "matte 0,0 replace" ${tmp}_label3d.png ${tmp}_label3d.miff
I also recommend changing the default fontscale to 1.0 (the easiest way to do this is using the FONTSCALE environmental variable, or you can edit the FONTSCALE setting in the label3d script)

Q: The molscript "auto" function doesn't do a good job assigning secondary structure. Can I fix it?

A: Yes. My favorite settings are:

set ss_hb on;
set energycutoff -1.0;
before I read in a PDB file. Those tend to lead to "sane" secondary structure assignments.

Q: Why does my raster3d rendering only show the back part of my surface?

A: this was a "feature" of raster3d - if the normals point what it considers the "wrong" way, they're not drawn. Upgrade to version 2.12 or better, which has the material declarations that should fix this.


Q: Why does my MSMS surface not work when I try to color it by potential or some other property?

A: Make sure you use the pdb_to_xyzrn script to generate the .vert and .face files (NOT pdb_to_xyzr) - since then the surface file will contain information about which residue/atom is closest to the generated vertex - which is what povscript+ uses to determine the vertex coloring scheme.


Q: Why do the atoms in my Raster3D output look different than my OpenGL/POVRay output?

A: This has to do with the "bsurf" setting. This has been changed as of povscript+ 2.02, which has the default "bsurf" setting at 0 (i.e. atoms are *NOT* rendered according to B factor).


Q: Why does the final image look "blocky"?

A: Chances are you're using only a few graphical segments to draw each object. I tend to use something like "set segments 10;" to get smooth transitions between objects (at least when I'm about to generate the final figure).


Q: I told POVScript+ to render my electron density at 1.0 sigma, but it doesn't look contoured properly.

A: make sure the map is normalized (just load the map in mapman and type "norm") before loading into povscript+, such that the map is scaled according to sigma levels.


Q: How do I show greek characters in POVRay?

A: Include the following in your .pov file:

#ifndef (GFont)
#declare GFont="symbol.ttf"
#end
(note that this will require you to have the symbol.ttf file, which POVRay does *not* include) Then find the text in the .povinc file you want to replace, and edit the entry to appear like the following:
text { ttf GFont, // <--- changed from RFont to GFont
  "a", FontWidth, 0
  scale <0.40, 0.40, 0.1>
  translate -y*0.15
  translate <-3.43, -3.80, 3.96>
}

Q: POVRay is spitting out an error like:

"#ifndef (1FILE_1 <----ERROR
Parse Error: Expected ')', undeclared identifier 'FILE_1' found instead"
whats wrong?

A: POVRay doesn't like variables that begin with numbers or contain math symbols (it will try to interpret them as numbers or math operations, not strings). So *don't* call your molscript input file something like "1mol.inp". "mol1.inp" is fine, however (if you do find yourself in this situation and feel lazy, something like
"sed 's:1FILE:FILE:g' 1file.povinc > file.povinc"
will quickly fix the problem).


Q: I can't see atoms, even though I told povscript+ to draw them as ball-and-stick or CPK!

A: Chances are, the atoms in the PDB either don't have B factors or are non-positive definite. Try "set bsurf 0;" in the povscript+ input file before drawing the atoms.



Q: I see the atoms, but they're not drawn with the atomradius I specify!

A: They're being rendered according to their B factors, not the van der Waal radius. Try "set bsurf 0;" to fix this.



Q: I see the atoms, but they're semi-transparent!

A: The default transparency for atoms rendered according to their B factors is 0.3 (30%) - this is so the principal axes are visible (if desired). "set btransp 0.0;" will turn this off.



Q: I can render my MSMS surface, but it doesn't look very good. What's up?

A: This is most likely due to a bug that existed in POVScript+ up through version 1.67. Upgrading should fix this. Another problem comes from a "bug" in MSMS surfaces, illustrated in the "BUGS" section of the manpage:

"Normal vectors of singular vertices point arbitrarily to the center of one of the probe they belong to."
Which doesn't help much in terms of visualization.


Q: I am trying to draw a ribbon diagram inside a semi-transparent surface. Although the surface is dimmed as if semi-transparent, I cannot see the protein.

A: You must draw non-transparent objects first (this is irrelevant for POVRay output, however). (thanks to David Cooper for this pointer)



Q: My electron density map is rendered, but it looks nothing like it does in O/XtalView/etc...! Is there something wrong?

A: My best guess is that you're using an older version of povscript (prior to 1.5) that uses the old marching tetrahedra algorithm, or using a DSN6 file with an older version of povscript (prior to 1.51) that didn't read the files properly. Try upgrading.



Q: Can I make the electron density look less cluttered?

A: Yes - try either the mapdepth (try something like -1) or mapthresh (try 0.5 or higher) variable. The mapdepth variable will step along the scalar field at coarser intervals, while the mapthresh variable will cluster vertices near gridpoints. The mapthresh setting seems to generate more esthetically pleasing results.



Q: How do I reference this piece of ju- err, program?

A: Feel free to either reference the webpage or the following:

Fenn, T.D., Ringe, D. and Petsko, G.A. (2003) "POVScript+: a program for model and data visualization using persistence of vision ray-tracing" J. Appl. Cryst. 36:944-947.


Go back to the povscript+ home page