VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Wed Feb 22 2012 - 11:12:27 CST

Hi,
  I have several comments on your question that you may find helpful.

On Tue, Feb 21, 2012 at 09:22:56PM -0200, Diego Enry Barreto Gomes wrote:
> Hi list,
>
> I was wondering if the developers could recommend a strategies to:
>
> 1) Render the frames
> - Which raytracer, can any of the available benefit from GPUs ?

Generally speaking, I always suggest that people start out by using
the Tachyon ray tracer that is built-into VMD. Of all of the ray tracing
options supported by VMD, Tachyon is the one that produces output that
most closely matches what you see in the VMD OpenGL window, and it is
the only one that supports some of the advanced shading features such
as angle-modulated transparency, outline shading, and other VMD-specific
shading features. The "TachyonInternal" renderer listed in the Renderer
menu is usually the best choice, because that version of Tachyon is actually
compiled into VMD itself, and it renders the VMD scene directly from RAM
without having to write the scene to a (potentially large) disk file first.
For very large structures, writing the scene files to disk can take a long
time, so I would recommend using TachyonInternal as your first choice.
Tachyon uses multi-core CPUs. By default it will use all of the CPU cores
on your machine. I haven't yet updated it to use GPU accelerated ray tracing,
but that is on my TODO list for the next year.

The NVIDIA gelato renderer is a RenderMan-like renderer that uses GPU
acceleration.

There are experimental versions of the Aqsis renderer floating around
that use GPU acceleration, but I haven't tried any of those yet.

> 2) Make the movie
> - Which codec to use.
> - Are there optimal resolutions ?
> - What kind of tricks can we use to reduce file size while keeping a crystal clear animation. (reduce color space ? remove background ? H/W ratio )

Generally speaking, the best advice for movie making is to avoid
drawing thin, high-contrast objects like 1-pixel wide lines.
Less saturated colors tend to survive compression better than
fully saturated colors. If you use a mid-level background (e.g. gray)
rather than black or white, you will get a better compression ratio.
If you render with antialiasing enabled, you will get less "pixel popping"
in your movies and the compression will be less damaging to the image
quality.

I suggest sticking with resolutions that are appropriate for
the movie encoder's internal bitrate. For the MPEG-1/MPEG-2
encoders used by the VMD movie maker, you really should
not go beyond 640x480 unless you change the encoder parameters
to increase the bitrate appropriately. If you use your own
encoder you can use any resolution you like, all the way up
to 4K resolution, but you'll have to run the movie encoding job
as a separate step with your own scripts. I would suggest that
you stick to a movie that is no higher resolution than the
on-screen window size that you intend to use when displaying
the movie in a presentation. Since many projectors are often
using resolutions of 1024x768 or so, this often means that
keeping the movie size down to DVD resolution (640x480 or
720x480) is often appropriate.

Cheers,
  John Stone
  johns_at_ks.uiuc.edu 

-- 
NIH Resource for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
http://www.ks.uiuc.edu/~johns/           Phone: 217-244-3349
http://www.ks.uiuc.edu/Research/vmd/       Fax: 217-244-6078