POV-Ray : Newsgroups : povray.general : How I did the censored frames : How I did the censored frames Server Time
9 Aug 2024 09:09:38 EDT (-0400)
  How I did the censored frames  
From: John VanSickle
Date: 4 Aug 2000 21:51:34
Message: <398B7380.9E4E45FB@erols.com>
Those of you who are curious how the blocky digital censoring was
done in my most recent IRTC entry, and haven't guessed, looky here:

The rendering was done in stages.

1) The uncensored scene is rendered normally, which for me is 320x240,
AA method 2, depth 2, threshold .1.  The frames are named srca####.tga,
where #### is the frame number.

2) The scene is rendered at 40x30 instead of 320x240, with anti-aliasing
method 2, recursion depth 3, threshold 0; this ensures lots of super-
sampling for each point.  These frames are named srcb####.tga.  The
camera has to be adjusted slightly so that the reduced version matches
the large version.  These frames take about two minutes each to render.

3) A simplified form of the scene is rendered at 40x30, in which the
object to be censored is pigmented full white, with ambient set to one
(or higher) and all other finish settings set to zero.  Everything that
might partially block the censored object is pigmented black.  The
camera is positioned exactly as in step 2.  Everything else is removed
from the scene, and the background set to black.  The frames are
rendered with the same AA settings as in step 2, but the rendering is
quite fast (<1 second per frame).  The resulting frames are named
mask####.tga.

4) tga2gif is used to transform the mask####.tga files into mask####.gif
files.  It turns out that the black background is assigned to color
index 0.  This makes the following step much easier.

5) A separate POV scene file uses the mask####.tga files as material
maps texturing a plane.  The camera is orthographic.  For index zero
(the black background), the material map uses an image map from file
srca####.tga.  For all other indices (any pixel that touches the object
being censored), srcb####.tga is used.  Anti-aliasing is not needed for
this step.

6) If the shot involves a fade or wipe from another shot, the resulting
frames are then used in another POV-scene to accomplish the fade.  If
not, the frames are done.

If I hadn't been able to pull this trick off, I probably would have
done an entirely different story idea altogether.

Regards,
John


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.