Instead of capturing an attribute for the dot product, you can also just use the "difference" and "hit normal" vectors. It simplifies things a bit. Also, you can apply a minimum distance threshold to avoid the pop-in near the camera that you see in 10:08 (this can be done with a switch node and the distance from the camera obviously). Hope it helps, that worked for me!
Ah great! Thanks, i also realosed more recently that you can instead compare the output normal to check if its equal to -1 on z and that also works! Too bad i figured it out too late for the vid ha
I just have a plane and the jumping point to make these extra nodes. I guess similar to what you have. When I create the grid node, it doesn't move to the camera position. It just stays on the ground and my initial nodes look just like yours except I don't know what's happening on the left side of your node tree.
Have you definetely added the "transform geometry" node after the grid and plugged in the location/rotation from the camera info node into the translation/rotation?
@AgustinCaniglia1992 Also, the camera cull seems to be locked to the object origin in the above tut. It's a weird dynamic of what the camera sees and how far you are from the object origin. And I don't think it can be changed. cgmatter's tut relies less on that I think.
Can't get the grid to work with the camera. I've put in transform geometry node but it isn't copying the location of the camera even though I thought I'd followed your steps exactly.