07 - Testing the New V-Ray Light Cache type "Hash Map"


Testing the New V-Ray Light Cache type "Hash Map"

Recently V-Ray Next updated to version 2.1. Besides other new features, one that caught my attention was the new Light Cache type called HASH MAP. 

 

I asked my friend Ricardo Ortiz - Chaos Group Product Specialist - and taught me how it works and I couldn't wait to try it by myself. 

I decided to use V-Ray Next GPU because its power against CPU is incredible and as you know, now V-Ray Next GPU is a full production engine with many capabilities. 

 

ABOUT HASH MAP:

Here's what you find in the V-Ray Documentation site:

"Hash map – This is the default light cache type for new scenes. It is a new light cache implementation that is simpler and produces less flickering in animations. Old light cache maps will not work with the new Hash method."

 

New Hash Map type.
New Hash Map type.

 

Yes... I know what you're thinking... What?

 

Well, I'm not a geek about raytracing calculations so here I share with you what Ricardo Ortiz says about the new features of V-Ray:

"In V-Ray Next we have implemented something we call Scene Intelligence, this feature of V-Ray involves tools that offer adaptability and automation in scenes, making everything easier and faster."

 

 

Many of these Scene Intelligence tools are based on the Light Cache calculation e.g. Adaptive Lights, Adaptive Dome, Auto Exposure, Auto White Balance. So in other words, now V-Ray is "smarter" and Light Cache is the way you get all of these new powerful tools.

 

THE TEST:

The scene test includes:

  • Animated lights: A V-Ray Sun with blue filter + 1 V-Ray Light in the lamp that changes color to make it more interesting to test.
  • Animated Geometry: A couple of modifiers over the couch and mirror to include geometry animation.
  • Animated Camera: The camera moves sightly to make it more complex.
  • V-Ray Dome: I chose an interior scene to make it more difficult to calculate, but I'm using the new Adaptive Dome that doesn't use portals anymore.
  • V-Ray Environment Fog: To get those volume lights that before V-Ray Next were infinite to render.
  • DOF in camera: The DOF and Bokeh were rendered from the camera.
  • G.I. solution: Brute Force + Light Cache

As you see many of these effects are hard to calculate for any renderer, plus, the animations in the scene were only possible to render before by using Brute Force + Brute Force because the old KD Tree version of Light Cache flickers a lot with them! 

 

Hardware used:

  • 1 - RTX 2080Ti
  • 1 - GTX 980

Render time:

  • 1:30 mins per frame with Noise threshold of 0.015 and Nvidia AI Denoiser

 

METHOD:

The method to use it is as simple as:

  1. Open your scene
  2. Choose V-Ray Next as renderer (no matters if you use GPU or CPU).
  3. Leave all the default settings as they are out of the box.
  4. Change Light Cache to WORLD because it's the best way to calculate for animations.
  5. Hit Render
  6. That's all!

It will calculate the Light Cache every frame, and I know that you thinking that it takes longer than precalculating 1 single pass of the Light Cache with Use Camera Path, but you don't have to compare it with that method because that way is used for animations where the camera moves but not the geometry of lights. This method has to be compared to a Brute Force + Brute Force workflow, so please believe me, IT'S FASTER!

I hope you like this post and also that you try this new and powerful workflow.

 

Thanks for Watching!

Mario Cameras

 

Share this article


Write a comment

Comments: 0