AWS Thinkbox Discussion Forums

Equirectangular image projection?

Can the Spherical mode of the ImageProjection node be used to project equirectangular spherical images, like you might get from a Nodal Ninja? If so what’s the trick?

This is what I get with a test sphere and spherical projector at the origin. Note that I’ve decreased the horizontal field of view from 360 so it’s easier to see that all of the image edges are converging at the pole. Back face culling is also off so we are looking at the outside of the sphere.

[attachment=1]equirectangular projection.PNG[/attachment]

This is what I was expecting to get using a lat-long equirectangular image with the top and bottom rows converging at their respective poles.

[attachment=0]expected_spherical.png[/attachment]
As I was writing this I just realized that maybe even though the docs say lat-long image perhaps the Spherical projector is actually expecting a environment map, i.e. mirror ball image? I’ll have to try reprojecting and see if that works better.

This is an old thread, but I’m curious if this is possible in Sequoia? Did you ever get it to work?

I created a GeoSphere in 3ds Max, exported as XMesh, and loaded with a Mesh Loader into Sequoia.
Then I created an Image Projection, picked a LatLong HDR shipping with an old version of 3ds Max, and projected in Panoramic mode with FOV of 360 degrees. I had to rotate the projector at 90 degrees about X to align it to the world correctly. I checked both Front and Back Projection to see the image on both sides of the mesh. I also switched the Mesh Loader’s display to not cull the backfaces so I can peek inside it.

The top panel shows the projected LatLong (equirectangular) image used in the projection.
The bottom left viewport shows the outside of the GeoSphere.
The bottom right viewport shows a view from the inside of the GeoSphere.

[attachment=0]sq_latlong_panoramic_on_a_sphere.png[/attachment]

Privacy | Site terms | Cookie preferences