transfer color from backplate onto points

[ranxerox]

hey all, I want to color points based on a backplate. I was thinking I could write a VOPSOP to transform the points into image space (based on a camera) and then use the P.x and P.y to access a texture. It almost works (by projecting into camera space), but not quite. Is there anything anybody can tell me here ?

thanks in advance

-ranxx

[mawi]

hey all, I want to color points based on a backplate. I was thinking I could write a VOPSOP to transform the points into image space (based on a camera) and then use the P.x and P.y to access a texture. It almost works (by projecting into camera space), but not quite. Is there anything anybody can tell me here ?

thanks in advance

-ranxx

Have a look at the toNDC function.

Transforms a position into normal device coordinates.

Contexts: image3d, chop, cop, pop, sop, surface, displace, fog, light, shadow, photon, cvex

1.

vector toNDC(vector )

2.

vector toNDC(string camera_name, vector point)

Transforms a position to the normal device coordinates for a camera. The point should be in the local space of the object (i.e. not in the space of the camera).

toNDC() will return values above and below the 0-1 range outside the view of the camera or light. To the right of the camera are values of 1+ and to the left are values of 0-. The same goes for the range above and below the camera or light.

Transforms a position into normal device coordinates. This space is only well-defined for the Displacement, Surface, and Light contexts.

[MrBag]

This might be old school and slow, but I tend to use a uvtexture SOP with projection from camera, and then the texni() expression in the point sop for color. To speed this up a VOP/vex sop makes a big difference. But sometimes quick proof of concept is more important

[ranxerox]

hey there, in looking at the Transform VOP I did notice you can use "space:ndc" as the destination space, but there doesn't seem to be a way to specify which camera's normalized device coordinates this refers to. I guess I can rewrite the transform VOP in vex and have it take a parameter to the camera, but it seems to me there should be a way to specify it on the transform VOP itself. I'm sure I'm missing something here.

thanks for your suggestion.

-ranxx

Have a look at the toNDC function.

[ranxerox]

thank you for this, I'm going to try this first, then try it in VEX (unless I can figure out how to do it in VOPS first).

-ranxx

This might be old school and slow, but I tend to use a uvtexture SOP with projection from camera, and then the texni() expression in the point sop for color. To speed this up a VOP/vex sop makes a big difference. But sometimes quick proof of concept is more important