Now that Proteopedia can run the signed applet (using a technique where you can load the page with a custom version of Jmol) I have been trying the different image outputs and come across some issues and hopes:
a) The current output of Jmol to VRML seems to have a problem around line 29 where the line beginning 'reset;center...' is not commented out and so causes nothing to load (in SwirlX3D viewer) or throws an error message about line 29 (vrml2pov converter). Either commenting out that line or hitting backspace to keep it running on with the line above allows both programs to work. b) Using VRML to convert to POV-Ray gives a high quality image (see , http://drop.io/85qrkhy/asset/1rpu-vrml-to-povray-jpg for example) except that regions of non-structured protein (loops) don't look smooth and nice like the rest of the image. Can this be fixed or improved somehow eventually? I looked at it directly in VRML and there it doesn't look good either. Using either means of rendering it looks like the loops consists of tiny multiple circles adjacent one another. The nucleic acid backbone looks beautifully very smooth though. Playing with lighting effects can produce even better results from Jmol via this route but the loops still detract. (Example of VRML to POV result with lighting effects added http://drop.io/85qrkhy/asset/1rpu-light-efffects-jpg . I think though I put the light in the wrong position in a cut and paste from another scene and got artifacts; however, it takes about an hour to render so I left it.) c) I am going from VRML to POV because the current implementation of Jmol's generated POV-Ray files does not create surface normals like Swiss-PDB Viewer does with smooth_triangles and Pymol does with a mesh2 objects. And thus it does not render a high quality image (see http://drop.io/85qrkhy/asset/1rpufromjmol-pov-jpg , for an example of the POV-Ray rendering as straight from Jmol)? In fact the Jmol image using Proteopedia's high quality toggle is much better, and unlike the files that come straight out of Pymol or Swiss-PDV viewer the objects are not suitable for being taken and used in advanced scenes that use fancy lighting effects or that have objects that do render at a very high resolution. The VRML output seems to contain the normals? Can the normals be generated by Jmol in its POV-Ray out so people can use that without needing the VRML to POV-Ray step (although I am not 100% sure the VRML output has it as I don't know VRML). More importantly, I think it would also make the POV-Ray output look better especially for people trying to make the rendering larger (although I had a hard time myself adjusting the height and width with Jmol's output to get a larger rendering where as I find it easily done with the output generated by Swiss PDB-viewer). Thanks, Wayne
------------------------------------------------------------------------------
_______________________________________________ Jmol-users mailing list Jmol-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jmol-users
