SCAD X The Mill - Rendering

To Compositing!


10/31/20233 min read

Addressing Texture & Material Issues

While developing the visuals of shot 03, Dee noticed that there were some issues with the plastic seat material that the boxes were sitting inside. The dark value of the seat made the scene appear a lot darker than we intended it to, and also, the reflectiveness of the seat made it appear like a solid plastic seat rather than a linoleum-coated one.

I noticed during my way figuring out Houdini Redshift, that Redshift tends to make the materials appear more shiny than how it is displayed in Substance Painter- even when used with the correct color space. Taking this into account, I fixed the textures and updated them- but also tweaked the material nodes of the plastic seat to make it appear closer to how we wanted.

This is how the scene looks before and after the improvement made. Lighting changes and the prop textures are the responsibility of Dee, and you can see her progress in her blog here.

Motion Blur on Shot 02

From the advice of Professor Fowler, me and Zach worked on achieving the motion blur that we wanted for this shot. The previous version, which was a temporary visual, had the motion blur applied to the entire shot as the camera was the main source of movement.

We thought that this is unrealistic, especially when the camera is moving alongside the front wheel in the same speed- thus, the elements that are to have motion blur would be the wheel and the ground.

Based on the suggestion of our professors, Zach worked on re-creating the velocity points of the wheel by emulating the circular motion of the tire, and then replacing it from the original geo so it is attached properly.

For the ground, he used a point wrangle node with VEX code to simulate a linear velocity point, as a regular trail node would give us a highly randomized result due to how the original geometries were scattered.

From what was provided, I tweaked the lighting, geometry, and blur settings to render the scene with proper AOVs. It was soon handed over to our compositor Mateo.

Set Dressing & Lighting on Shot 06

Based on the set dressing Mateo added on, I polished the lighting and materials to better match with the other shots. We have encountered some technical issues- such as grass jittering in the final render, and Redshift being unable to calculate the point clouds for the rock geometry.

  1. Grass & Small Rocks Jittering

  • (I currently do not have access to the render farm files due to the remote connection being unstable- I will update a video example of what it looks like.)

  • It was the same issue that Mateo encountered last week for his render of shot 05. This was an easy fix, as it was eliminated once I cached out the scattered geometries.

  1. Unable to calculate the point clouds for the rock geo

  • I asked Professor Fowler about this issue, and it was solved by changing the calculation method from Redshift Point Clouds to Redshift Instances.

  • This, from our guess, reduces the memory usage of the program trying to re-load every single geometry every time each device in the render farm opens the file to render. I plan to search for more documentations that might explain why this happens, and which method of calculation I should use by case.

To-Do for Week 8

  • Render Shot 2 with Volumetric Light

  • Render Shot 6 with Volumetric Light

  • Final Iteration of Tuk-tuk Lookdev (Focusing on Interior)