From: John Stone (
Date: Wed Mar 21 2018 - 10:36:54 CDT

There isn't a way to do this at present, but I've been pondering
adding a global scaling factor for lines. Lines are always a
problem for ray tracing engines since they have no "size" per se.

The only reason VMD even uses lines is that they are very very
low cost for a GPU to draw.

I think that there are two solutions I would like to pursue:
 1) Add a global scaling factor for lines when exported to ray tracing
    engines so that this can be adjusted. This should be relatively
    easy to implement and then you could override it at runtime
    however you like. I might be able to get this implemented
    in the next week or so.

 2) Have cylinderical variants of everything that VMD would normally
    try to draw with lines. This is not quite as easy to implement
    since connected lines are simpler to draw than connected cylinders,
    but VMD already has the mechanism for most of what's needed here,
    and the affected representations (any/all "wireframe" modes) aren't
    too numerous, so this can probably be done before VMD 1.9.4 is released.

If you're willing to try a test version of VMD, it would be pretty easy
for me to come up with an initial implementation of 1) and have you test it.

  John Stone

On Wed, Mar 21, 2018 at 12:39:19PM +0000, Rune Thomas Kidmose wrote:
> Hi,
> Is there a way to reduce the thickness of wire frames on an isosurface
> when rendering with TachyonL?
> I have tried setting the "size" parameter to sub 1 via the below tcl
> command.
> mol modstyle 0 1 Isosurface 0.053959 0 0 1 1 0.3
> however it seems 1 is the lowest it can go.
> My problem is when I want to render figures using TachyonL,
> of close-up views of density displayed as a wire frame isosurface. The
> lowest "size" setting of the wire frame, still generates far to thick
> wire frames to enable viewing the model inside the density.
> You can see an example image of the problem below (was not sure I could
> attach files to this mail)
> [1]
> The above image is rendered with TachyonL-optix, but the same applies to
> -OSPRay
> This is how the model and isosurface looks in the display window
> pre-rendering:
> [2]
> And this is the results using "just" tachyon rendering:
> [3]
> where there is no problem with the wire frame thickness.
> As I would really like to use TachyonL-optix, how can I solve this
> problem when rendering this kind of closeups of wire frame isosurfaces?
> Cheers
> Ruki
> References
> Visible links
> 1.
> 2.
> 3.

NIH Center for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801           Phone: 217-244-3349