Steve Bevins

Edit: In classic fashion I found a non hacky fix right after posting this. Just switch the output type from packed raster to deepRaster and all the tiled settings work as expected.

Hey Simon, just confirming that its definitely broken(to my knowledge), but theres a hacky fix for it.

Hacky Fix: You can put a tile split node before the heightfield output, set the node to output single tiles, and the tile number to $F-1. You can then tell the heightfield output node export a sequence, which is how many tiles you have. So I normally expression link the max frames to the heightfieldTile tilex parm * heightfieldTile tiley parm

Why I think its broken: If you actually unlock the heightfield output node and look at how its exporting, the COP network is pointing at a window node thats hooked up to invisible parms (and somehow this is working, but not iterating) and the ropOut in that network is not iterating at all, because it too is hooked to inivisible parms that aren't working. Seems like it started to get implemented but it just never got finished. Or I'm using it completely wrong.

I totally agree. Really fantastic work man, it all looks great!

@mottchzee

What I would recommend doing is caching out your rbd simulation of the sphere breaking through the wall.

You most definitely need to have the pyro be tracking the post-simulated sphere, because like you said before it gets run through the solver its just a static object with initial velocities. The reason I recommend caching your sim is that if it is not cached every frame that the pyro needs to track the sphere, it will have to run that rbd sim first. If you cache the sim to disk (using a ropout or file node in conjunction with dopImport) houdini only has to look up point positions, and not run a full rbd sim.

If it's still not working after doing that feel free to throw a hip file up and i'll check it out for you.