With the new update to Vray for Rhino I (finally) discoverd Chaos Vantage (Chaos’s own Unreal?).
My main question is: Why don’t we have such fast preview options in Vray for Rhino? I’ve got an RTX3080 and the realtime options (CPU/CUDA/RTX) in Vray for Rhino take many seconds to create something only half way as good as what Vantage creates at 60fps!
I mean I don’t want it for the final render - as Vantage is clearly not up to much when it comes to detailed bump maps and other complex textures - But it’d be nice for initial setup work - especially lighing etc.
Chaos Vantage is built from scratch for fast previews in higher FPS, it relies on dedicated raytracing hardware in the latest Nvidia Turing and Ampere GPUs.
On the other hand, V-Ray GPU(offline rendering) offers high depth for Global illumination/Reflection/Refraction, and advanced shading tools and features, some of which can be expensive on real-time performance and remain exclusive to V-Ray and V-Ray GPU
Chaos Vantage and V-Ray GPU serve different production needs, and in some cases could be used together in the same rendering pipeline.
We are looking into ways to improve V-Ray GPU’s interactivity for look-development and lighting, it is high priority to the development team. we have GPU Light Cache working in IPR in internal builds, this should come out soon, and should give a good performance boost, specially for interior scenes
Glad to hear the feedback on update 2
I came from Keyshot as well, I worked at high-end print for the past 6 years as a concept artist and visualizer, mostly automotive and product visualization. I transitioned fully to V-Ray GPU around 2016-2017, mostly for shading quality, and production rendering speed
GPU LC has helped massively with interactivity. It is working in update 2 in 3Ds Max, and hopefully soon for other V-Ray products. There are other plans to improve interactivity in V-Ray GPU, it is something I discuss regularly with development teams
Chaos Vantage is progressing fast as well, the team is focused on quality and shading features at the meantime
Support will help with the Vantage issue, and follow along with you. And if you have questions or feedback on the GPU workflow feel free to reach out to me.
Thanks for checking, this is part the team is working on.. Having LC work in IPR, it should boost interactivity for look-development and lighting
And it should make the scene much closer to the final render on both speed and quality, with 100 GI bounces
Forgive me for asking, Muhammed, but I have a question with regards to LC on the GPU: Why should it boost interactivity? In my experience every time I change anything when using GPU IPR with LC, it has to recalculated the Light Cache, thus making interactivity slower, not faster. Or is there something I’m missing? I appreciate that overall BF is slower to get to the final image, but for pure interactivity it seems faster to me.
This behavior is mostly a bug, we have a fix coming soon. Basically undersampling settings get ignored, and there is like a short delay when you make any change
LC uses 100 GI bounces, and Adaptive DomeLight needs it to work
If you are using an HDRI to light up an interior or studio scene, with LC you get less noise and much faster previews overall.
Using BF/BF in this situation you get a lot more noise, and lighting will look different as adaptive domelight is not gonna work correctly, and you have less bouncing of light in the scene so you get a darker result than expected
If you try to use more bounces with BF/BF it hurts the performance as well.
On another note, I talk to the GPU team every day about GPU LC. We have plans to improve interactivity, so this first pass should be faster to calculate.
And we are working on improving Undersampling, depending on render resolution. There is already something in place for CPU rendering, we are looking into this now
GPU LC is progressing fast on both speed and time to first pixel. I will update the thread when we have something in the nightlies
