Basically as the title states. I’ve been seeing a lot of footage come out of Pyro 4 with its new clouds, and whilst watching this clip, I noticed the reflections on the water were disconnected from the reality of what was going on above it.

Noteably:

– The water not reflecting the clouds above.

– The delay between shifting reflection maps.

Now I know the currently implementation is more of a bake once implementation, but we saw at Citcon the work done on GI (for opaque surfaces), transparent GI (for non-opaque surfaces) and glossy GI (for reflective opaque surfaces), and I was wondering if this work will allow for more accurate reflections on the water’s surface for things like the clouds in the sky and the rings in orbit?

Thanks.

And also a side note, the rings rendering over the clouds. Just thought I’d point it out to make an already amazing looking scene look even better.

Hi,

In the above video we should have been seeing screen space reflections for the sky, but these are oddly missing (I’ll get someone to check why, assuming it’s not just user options). Instead what we are seeing is the cubemap reflections, which are regularly re-baked on movement or sun angle changes, but currently aren’t re-baked with the movement of the clouds as this has only recently been enabled (unexpectedly early!) and we haven’t had a chance to improve the cubemap capture system.

As for GI, this is mostly about bouncing diffuse (soft) lighting rather than specular (shiny) reflections, but we are upgrading the reflections at the same time in order to avoid a disconnect between the diffuse and specular bounced light. However when it comes to the clouds, they won’t be ray-traced (cost would be insane), and any ray-traced reflections that miss opaque geometry will fall back to using either screen space reflections or a cubemap, so again the solution will be to bake the cubemaps much more frequently with moving clouds.

Ali Brown

Senior Director of Graphics & Procedural Tech