From: Charles McAnany (vmdnamd_at_charlesmcanany.com)
Date: Wed Jul 29 2015 - 12:04:59 CDT

John,
Thanks for the reply. I'm not quite sure what the screen height and screen
distance are changing (I gather they're terms from the rendering community
and not the photography community), but they produce the effect I want. Out
of curiosity, what does the 'focal length' parameter control? It doesn't
seem to do anything for me.

Is there a way to have mouseover text in Tk? So, when the mouse hovers over
'Screen Hgt', help pops up? I'd be happy to help contribute some text if it
is possible.

Regarding tilt, your idea of by-atom focus is much better than tilt-shift
since it's so general. Heck, even something as simple as focusby beta would
be useful for visualizing pdb structures and it'd be completely impossible
with tilt-shift. I look forward to playing with this feature when it comes
out!

Cheers,
Charles McAnany

On Tue, Jul 28, 2015 at 1:53 PM, John Stone <johns_at_ks.uiuc.edu> wrote:

> Hi,
> I'm on vacation and can't write a long response, but you can
> control focal length indirectly by changing the display settings
> camera distance, screen height, and screen width, which are effectively
> the same parameters that would be computed if you used a focal length or
> field-of-view type of control instead. If you want a longer focal length
> and narrower field of view (for less "eggy" looking spheres in the
> corners),
> just change the screen height to be smaller, or change the screen distance
> to be a larger negative value. You can compute these from a field of view
> angle with a couple of lines of tcl.
>
> I understand your request for tilt-shift type DoF optics.
> At present there's no ability to tilt the focal plane, and although I'm
> very familiar with tilt-shift lenses and similar optics, I think that
> it would require a lot of work for end users to learn how to use this
> effectively. Instead, I have been planning on implementing a
> different scheme whereby the user will provide an atom selection or a
> volumetric map that defines regions of crisp focus, and the rendering
> code will use this to drive the focal blur effect. This would solve
> the problem of creating regions of crisp focus that are unrelated to
> the particular shape of the focal plane. The tilt-shift scheme would
> work great for a traditional camera, but it isn't viable for things
> like panoramic projections or spherical cameras, which are already
> implemented in the new developmental version of VMD 1.9.3. This is
> one of my motivations for considering a very different approach.
> I think that users will find a selection-based focal blur scheme more
> usable without having to become optical experts, but I'm certainly
> willing to listen to what other people think about this too.
> Writing the rendering code for a tilt-shift camera lens is easy, but
> doing the user interfaces to adjust it and make it usable by non-experts
> is likely to be substantially harder. Let me know what you think.
>
> Cheers,
> John Stone
>
>
> On Mon, Jul 27, 2015 at 12:59:29PM -0400, Charles McAnany wrote:
> > Friends,
> > The photographer in me has been having too much fun with the DOF
> features
> > in the lastest VMD, and it got me thinking about the camera. There
> are two
> > things I haven't been able to figure out how to do:
> >
> > 1. Change the focal length of the camera. As it stands, the focal
> lengths
> > available are 'perspective' and 'orthographic'. Sometimes, the
> > 'perspective' is too wide-angle and the molecule gets heavily
> distorted
> > when I fill the screen with it. The orthographic projection, however,
> is
> > too flat and doesn't convey the sense of depth I want (and also I
> can't
> > use DOF with the orthographic projection). It seems that zooming in
> VMD is
> > similar to moving the molecule relative to the camera, so the only
> way to
> > get the perspective I want is to push the molecule away, then render,
> then
> > crop the rendered image.
> > If it's not currently possible, I think a simple slider in the display
> > settings box that lets the user select a focal length (maybe a slider
> from
> > 'perspective' to 'orthographic' to keep it straightforward) would be
> very
> > handy for many users.
> >
> > 2. Tilt the lens relative to the film plane (see
> > [1]https://en.wikipedia.org/wiki/Tilt%E2%80%93shift_photography). For
> > example, yesterday I was rendering a dimer and I wanted the interface
> to
> > be in focus. However, with the current DOF tool, I couldn't figure
> out a
> > way to skew the plane of focus to move it where I want. This is, to be
> > sure, an unusual need, and if it'd be hard to implement in Tachyon
> then I
> > totally understand.
> > I can't think of a time when I've wanted to shift the lens when
> rendering
> > proteins, but, again, if it's easy to implement, then it'd be a cool
> > feature to have.
> >
> > Cheers,
> > Charles.
> >
> > References
> >
> > Visible links
> > 1. https://en.wikipedia.org/wiki/Tilt%E2%80%93shift_photography
>
> --
> NIH Center 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/
>