|
|
alex wrote:
>
> I think that you cannot specify a non-perpendicular direction vector using
> look_at *and* direction (I don't have sources handy to check that), but using
> transform you do and that's why I didn't get different result varying the
> direction vector.
>
> The camera specification I was using was:
> camera {
> perspective //keyword is facultative in this case
> location <50,10,30>
> direction vnormalize(Direction)
> look_at <0,10,0>
> blur_samples 120 //add samples if you got a fast computer
> aperture 10
> focal_point <5,10,5>
> }
> (from focalblur.pov)
>
> whatever Direction is, the plane of focus does not change.
>
> But if I use this specification:
> #declare S_angle=(CamLook.y-CamLoc.y)/CamLoc.z;
>
> #declare Shear= transform {
> matrix < 1, 0, 0,
> 0, 1, -S_angle,
> 0, 0, 1,
> 0, 0, 0 >
> }
> camera {
> perspective //keyword is facultative in this case
> location CamLoc
> angle 90
> transform Shear // comment out to see 'falling buildings'
> rotate <0,45,0>
> look_at CamLook
>
> blur_samples 120 //add samples if you got a fast computer
> aperture 10
> focal_point CamLook
> }
> (from shear.pov)
>
> the focal plane is *not* parallel to the image plane.
> Just add focal blur to the shear.pov example scene and see for yourself.
>
> I think that the focal plane is (correctly) parallel to the image plane only
> when the camera vectors are perpendicular.
>
> Exactly what I was looking for. :)
>
> Now, all I need is a ton of help in working out the transormations needed to
> emulate a tilt/shift lens....any volunteers?
>
> --alex
>
>
Greetings, all. I have a macro that simulates a shifted lens which may
help you. I will repost it in p.b.s-f. as FieldCam.zip. In the zip are
three folders; one for Mac users, one for Unix users and one for M$
users. Hope it helps
John
--
I will be brief but not nearly so brief as Salvador Dali, who gave the
world's shortest speech. He said, "I will be so brief I am already
finished," then he sat down.
Post a reply to this message
|
|