John  has made some great animations where the plot line is advanced by
overlaying text ontop of the image.

When I wanted to do this for a much earlier submission, I ran into a
roadblock given my interpretation of the IRTC rules.  Instead of using a
paint or video editing program, I spent weeks developing a way to do
this in povray.  Mine looked much sloppier than those in John's
animations, such that I wonder if he used a different technique.

Does IRTC care about text overlays done by a video program?  Again, I'm
not whining about anything, just wondering what I can get away with next
time!

Post a reply to this message

From stuff that is found in the ng, a macro was made that can keep text
in front of the camera. That might help. See     Re: OT: e-mail redirecting
in povray.general, dated 8/5/99 2:22 PM

"Greg M. Johnson" wrote:

> John  has made some great animations where the plot line is advanced by
> overlaying text ontop of the image.
>
> When I wanted to do this for a much earlier submission, I ran into a
> roadblock given my interpretation of the IRTC rules.  Instead of using a
> paint or video editing program, I spent weeks developing a way to do
> this in povray.  Mine looked much sloppier than those in John's
> animations, such that I wonder if he used a different technique.
>
> Does IRTC care about text overlays done by a video program?  Again, I'm
> not whining about anything, just wondering what I can get away with next
> time!

Post a reply to this message

Greg M. Johnson wrote:
> 
> John  has made some great animations where the plot line is advanced
> by overlaying text ontop of the image.
> 
> When I wanted to do this for a much earlier submission, I ran into a
> roadblock given my interpretation of the IRTC rules.  Instead of using
> a paint or video editing program, I spent weeks developing a way to do
> this in povray.  Mine looked much sloppier than those in John's
> animations, such that I wonder if he used a different technique.
> 
> Does IRTC care about text overlays done by a video program?  Again,
> I'm not whining about anything, just wondering what I can get away
> with next time!

The subtitles are done in straight POV scene code.  If you download the
source archive for my latest entry (boxer.zip), you'll find in it a
series of files named bsp_t*.inc.  These files place the objects
relative to the camera's viewpoint, so that they are in the camera's
view.

Prior to invoking these files, I have five values defined:

* CamU:  The up vector of the defined camera.
* CamR:  The right vector of the defined camera.
* CamD:  The direction vector of the defined camera.
* CamL:  The location vector of the defined camera.
* CamZ:  The zoom length of the camera.

I then place the text as if the camera were at the origin, looking
in the positive z direction, with all of the objects in the z=CamZ
plane.  I then apply a transformation matrix to the objects to place
them in front of the camera.

As a rule I make the text white, and behind it I put a filtering
polygon to provide contrast between the text and the background.
In the animation "Death to Rusty!" I also used green lettering that
resembled an LED display.

Your attempts at the same effect used red lettering.  The manner in
which MPEG compresses color causes the edges of bright red objects to be
jagged.  You can see this effect in the titling shot of "Death to
Rusty!" The frames were rendered with full anti-aliasing, and the edges
were perfectly smooth before the MPEG compiler got a hold of them; after
that they were jagged.

Hope this helps,
John

Post a reply to this message