Render Phoenix Fire and Smoke with VRayNext GPU

[TubeSmokeGuy]

Hello guys,

i got an issue rendering Phoenix Fire and Smoke with VRayNextGPU. In legacy VRay and VRayNext CPU everything works fine.

Just created a simple simulation for testing. Please have a look at my attachments and pay attention to the image numbers and the areas which i highlighted in red boxes.
First I rendered with "Create Fire Lights" turned on in Phoenix Volumetric Render settings. Then i got the warning message "Unsupported: Create Fire Lights ...", so I tunred it off to get rid of the warning message.
But see for yourselfs ...

01: Turned off "Create Fire Lights" in Phoenix Volumetric Render settings. Nevertheless - warning message appears in VRay Log. See also the red marked entry in VRay Log which says, that the adaptive lights could not be calculated during the prepass for LightCache - so rendertimes become crazy high for this small image by calculation at runtime.
02: Turned off the complete fire. Nevertheless - warning message appears in VRay Log. See also the red marked entry in VRay Log which says, that the adaptive lights could not be calculated during the prepass for LightCache - so rendertimes become crazy high for this small image by calculation at runtime.
03: Just replaced the PhoenixSimulator with a VRayEnvironmentFog BoxGizmo for checking out if it is a general problem with volumetric fluid effects. But as you see there is no warning message concerning "Fire Lights" and the adaptive lights data has been successfully written to the LightCache during prepass which lowers the rendertime critically. And please have a look at the red marked area in the rendering - there are crazy artifacts in the fog and the space where the fog is interacting with light, reflection and solid object. In CPU mode the fog renders finde and the render looks pretty.

So my question is: What can I do when i want to render the Phoenix Effects with VRayNext GPU in short rendertimes and without warnings? What can I do to render the Fog scene without artifacts? Will there be a fix in the future?

Thanks in advance!

[georgi.zhekov]

Hello,

The warnings are just that - they are not something that's a major issue in a scene. It's just useful info on what is happening during the render, but nothing to worry about.

Indeed the warning for the Create Fire Lights option is always showing no matter if the option is on or off so we will fix that.

As for the Adaptive lights message - it's not that they won't work with the Volume Grid - they will, it's just that they will be computed during the rendering and not as a prepass. So the render should start a bit slower and speedup when the adaptive lights kick in.
Note that comparing Environment fog and VolumeGrid will not produce the same result so their render times will be different.

The artifacts with the env fog shouldn't be happening though. Can you send us a scene so we can check what is going on?

Thank you,

[TubeSmokeGuy]

Great, thanks for that info.

Will the fire lights be supported by GPU somewhere in the future?

Fog Scene attached.

[matanov]

Will the fire lights be supported by GPU somewhere in the future?

Yes.

The artifacts in the fog scene are caused by the color mapping settings. This will be fixed, but my question is why have set those like that?

[TubeSmokeGuy]

I did those settings cause it is or was my or our most proven standard. But after your question I checked the documentation and found the advice in brown color here https://docs.chaosgroup.com/display/...+Color+Mapping , which says, that the subpixel mapping lead to that kind of artifacts. So sorry for poking you but that was not obvious to me. In general I just wanted to check how much work it will be for us to switch to or expand our pipeline for GPU rendering, leaving as much of our vray projects, settings, assets an stuff as they are, what we have to notice and how many changes do we have to do manually to keep or get beautyful results as we got them with legacy or Next CPU versions. On the other hand i would like to get an idea of how useful GPU rendering is for us at that point. Maybe you noticed my other topic with the EdgeTex in this forum, this was even a part of this exploration process. All in all i just had thrown a bunch of our project scenes into the "GPU render pod" and then pressed the render button, facing all messages and warnings. Then I did everything to prepare the scenes for rendering without errors and warnings by paying attention to this page https://docs.chaosgroup.com/display/...orted+Features . In most cases it worked quickly and without any problems. Maybe you can add the subpixel topic there and give a hint to the "VRayEdgeTex" entry with the "VRayBumpMaterial"-workaround you gave me, because it is just marked with a normal hook and without a footnote so that I thought I can use it as usual.

So long, thank you!

[matanov]

Ok, thanks for the feedback. Generally, the default settings should be good enough for almost everything. If you come across any other bump on your way to GPU rendering - don't hesitate to ask for help