"Kenneth" <kdw### [at] gmailcom> wrote:
> Jaime Vives Piqueres <jai### [at] ignoranciaorg> wrote:
>
> >
> >    Well, I was thinking to just use it for the first CG rendering, not
> > for the final composite one... so you wouldn't need to do the
> > undistort-redistort trick.
> >
>
> YES, that makes more sense; it eliminates some unnecessary steps. Thanks for the
> tip! I should have thought of it myself :-/

So, as I currently understand the mesh camera, what I would need to do is make
an *almost* flat mesh, but with some slight (radial) curvature going out toward
the edges (the amount of curvature 'matching' the lens distortions(s) of my
video camera), then use that mesh *as* the camera for my CG-rendering. Does that
sound logical and correct?

Then the only remaining step is to composite the final images, a simple
procedure!

Post a reply to this message

Alain <kua### [at] videotronca> wrote:

>
> Spherical aberation cause a coma effect. Bright points at some distance
> from the axis tend to show some egg or plume shaped fuzzyness oriented
> radialy arount the center point of the image.
>

....and also 'purple fringing' toward the edges. (It's interesting that the
cause of *that* is not well-understood, based on the research I've done.) My
Canon camera does show this, unfortunately.  BTW, I seem to recall that one
of the newsgroup members here-- I don't remember who it was-- came up with some
code to actually *create* this purple fringing, during POV-Ray renders (the
purpose being to make our CG renders look more like real photos.) That extra
bit of realism-- well, 'lens realism'-- would be useful for my matchmoving
scheme.

>
> Barreling can be simulated using the spherical or ultra whide angle POV
> cameras.
>

Hmm, that *is* true. I should experiment with that idea.

Post a reply to this message

"Kenneth" <kdw### [at] gmailcom> wrote:

>
> So, as I currently understand the mesh camera, what I would need to do is make
> an *almost* flat mesh, but with some slight (radial) curvature going out toward
> the edges (the amount of curvature 'matching' the lens distortions(s) of my
> video camera), then use that mesh *as* the camera for my CG-rendering.

I just thought of another, more simplistic way to duplicate the video camera's
lens distortion, without resorting to the mesh camera: POV-Ray's perspective
camera can have a 'normal' pattern applied to it, so that the resulting render
is distorted (as, for example, using a ripples pattern to give the image a wavy
'underwater' appearance.) With a properly-made bump_map (possibly from a
mathematical function) I might be able to duplicate that distortion during the
CG renders.

Post a reply to this message

> So, as I currently understand the mesh camera, what I would need to do is make
> an *almost* flat mesh, but with some slight (radial) curvature going out toward
> the edges (the amount of curvature 'matching' the lens distortions(s) of my
> video camera), then use that mesh *as* the camera for my CG-rendering. Does that
> sound logical and correct?

   Yes, that would be it...

--
Jaime

Post a reply to this message

> "Kenneth" <kdw### [at] gmailcom> wrote: I just thought of another,
> more simplistic way to duplicate the video camera's lens distortion,
> without resorting to the mesh camera: POV-Ray's perspective camera
> can have a 'normal' pattern applied to it, so that the resulting
> render is distorted (as, for example, using a ripples pattern to give
> the image a wavy 'underwater' appearance.) With a properly-made
> bump_map (possibly from a mathematical function) I might be able to
> duplicate that distortion during the CG renders.

   Yes, that would work too, and would be a lot easier to setup, indeed.
In fact, before the meshcam feature came out, I sometimes used this
approach to simulate barreling.

--
Jaime

Post a reply to this message