Jaime Vives Piqueres wrote:

> <...> I think you are going the hard way... I prefer to think that up and
> right
> have nothing to do at all with camera position or direction, but rather
> with the "film" or "visualization window" size. I prefer to have it fixed
> at the default orientation, only sizing them for ratio purposes, and then
> using only location, direction (some prefer angle) and look_at to control
> the camera. <...>

You are right. It is the hard way. Using the "position" and "look_at" vectors
is much simpler to imagine. But how do you work around the jump in orientation
from
camera { orthographic location <0,1,0.01> look_at <0,0,0> }
to
camera { orthographic location <0,1,0.00> look_at <0,0,0> }
?
That is the only reason I chose "up" and "right": to avoid discontinuities...
when the
camera is pointing in directions close to the y-direction. I'd gladly switch to

"position" and "look_at" if there was an easy way to get
round the jump (perhaps with angle? but when will the camera flip round?).

Regards,
Ralph

Post a reply to this message