Re: Computer Compositing

[T3D Larry Berlin <lberlin@xxxxxxxxx>]

Bill Stratemeyer  writes:
> I have uploaded a modified deviation test called devt.jpg
>...........................
>> T3D john bercovitz wrote:
>> There's too little difference there for me to see any distortions.
>> Sometimes I think I can imagine that one end or another of the
>> brick (rectangular parallelpiped 8-) is larger or smaller than the
>> other end of the same brick but it could be just my imagination.
>> The thing is, I _know_ it's a brick and it's hard to clear my
>> preconceptions.
>> 
>> One comment is that in a crosseyed view, you get keystoning for
>> sure.  So it's better to have infinity at 65 mm separation if you
>> can view it like that.

Bill Stratemeyer  writes:
>  I will redo a new test with more parallax.  And set the infinity
>spacing to 65 mm.  Maybe now just 2 objects one wire frame and one
>solid.  Say the Brick, both objects in the same orientation viewed
>square
>on and several degrees above. Or would it be preferable to modify
>the "rectangular parallelpiped 8-)" object so that there are no right
>angles.

I like Bill's illustration. It does a great job showing that the object
(intrinsic) parallax  stays relatively the same for placement at different
depths (extrinsic parallax). There is less distortion to the objects placed
at the different depths than occurs overall when I change my viewing
distance from the monitor (from relatively close to about 6 feet away).

John commented that there was keystoning in the crossed view but I'm not
sure where to observe it in this illustration. Maybe that will be clear by
comparison in the next version of the image?

Larry Berlin


------------------------------

End of TECH3D Digest 44
***********************