I am familiar with how to create an HDRI image of a set using a mirror ball to integrate cg elements with the environment but how do you do this if say a cg character is going to be running down a long street with the camera tracking alongside? What is the workflow for something like that? Do you chain several HDRI images created at different intervals along the way?
Thanks in advance for any guidance!
You could, but for best results try making a low-poly version of your street environment and projecting or otherwise texturing your HDRI images onto that. Then just use regular GI techniques to light the character from the environment, using HDRI lighting for the sky only.
Is the method you described the typical workflow in larger films these days? Bad example but say in like the first transformers film where you have robots battling down the highway, would you photograph all along the highway and project that onto a low res set to create the correct lighting on the robots?
Also, if this technique is being used with controlled lighting like with an interior would you want to photograph the set with the lighting set to how its going to be filmed or would you photograph the set with bright lights so its all flat where you would then create your own CG lighting/shadows to match?
Thanks in advance!
It depends. Different vfx studios do things slightly differently.
On one film which had CG cars on photographed background plates I used a combination of static and moving fisheye panoramas along with reflection cards with reference photography and the plate projected onto them.
The static fisheye panoramas were shot as regular HDR images. The moving fisheye panoramas however were shot from a moving car with two motion picture cameras - one facing forward, the other facing backwards. They were shot on film and exposed the same as the hero shot (so they had the same dynamics range) - I unwrapped both fisheyes into a spherical format and composited them together, blending and blurring the stretching to cover up the missing coverage (from the sides mainly) in the panorama. Once the hero shot was matchmoved I constrained my environment light to the matchmoved camera.
I then projected parts of the plate onto reflection cards (could have done this onto lowres geo, but reflection cards rendered quicker). For other shots with passing cars, I projected reference photos of the cars onto cards which I then animated going past the CG cars.
One thing I haven’t gotten around to trying is using photogrammetry to create HDR geo (then rendered to a pointcloud/brickmap) for environment lighting.
This paper talks about how ILM handle physical lighting and shading - it also covers how they shoot light probes and HDRs on set - most of what they do is fairly standard practice across the industry.
http://renderwonk.com/publications/s2010-shading-course/snow/sigg2010_physhadcourse_ILM.pdf
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.