(Moved from support channel to bugs)
Hello,
I have been doing some tests with the 6 point normals and im getting some odd results. The angles seem to be incorrect and it looks like the lights are coming from an angle instead of directly along the camera xyz axis. As you can imagine, its causing a lot of headaches in my workflow.
I set up a simple test to confirm. I made a sphere with smoke. No motion or anything. I dont have any weird rotations on my camera, it is aligned perfectly with the x axis of the simulation with z pointing up. I have created normals using the 6 point normals as well as the density gradient normals and the results are very different. Here are some images…
Density gradient normals, red, green and blue channels.
6 Point normals, red, green, blue channels.
As you can see channels are totally messed up. The red and green channels are very similar and they are pointing from a 45 degree angle. The blue is all over the place. This is with normalization turned on. I rendered it at 1 gamma so its easier to see the gradients.
Here is a test with a red light on the right and a green light on the bottom and blue along camera vector, which is half of the 6 point lighting setup. As you can see, the lights are perfectly aligned along their respective axis and looks similar to the density normals.
Is this a problem with the software or could this be a problem with my install? Are you able to repro this issue on your end? The best way to test is just render a simple sphere. Im on version 1.2.4
Thanks for any help!