arrow back
back 

When two 3D students revolutionise the production workflow


News . 10 May. 2024
Cycle Pro Animation 3D & Effets Spéciaux / Toulouse

Adrien Lhabitant and Tom Perony, students in their 4th year of the Animated Film course at the Toulouse campus, are the perfect example of the fusion between artistic creativity and technical expertise : they have created a tool to exploit the powerful AI denoiser used by Pixar Animation Studios. 

Currently working on their final year film, Adrien and Tom have also undertaken the development of an innovative tool aimed at exploiting the full potential of Pixar’s rendering engine denoising system, RenderMan.

This system offers a valuable advantage by considerably reducing the time needed for the rendering process (for example, reducing the noise present in rendered images : what can appear as unwanted grains or imperfections, often due to technical limitations). This solution therefore optimises resources and maximises the efficiency of the production workflow.

To learn a little more about their backgrounds and their approach, they answer a few questions : 

Can you tell us about your career as ESMA and where you are today ? 

[A] We both started at ESMA in 2019 with the Entertainment preparatory course, and then decided to follow the 3D Animated Film course. In the thirb year, we chose to specialise in rendering and, at the same time, developed skills related to the technical aspect of 3D production. These include the production pipeline, tool development and solving technical problems.

You are working on the graduation films El DODORADO and RUSH MORE ; what positions do you hold ? 

[T] Like Adrien on EL DODORADO, on RUSH MORE I’m Technical Director (TD) as well as Fx artist.

So we both have to develop solutions to facilitate and automate the production process. We have worked together on a number of tools, in particular Dopoise, to address similar issues. 

Can you tell us more about the creation of your tool ?  

[T] Dopoise os a tool thzt allows Renderma,’s new IA denoiser to be used efficiently in an production context, in particular by allowing it to be used on a Tractor renderfarm. We have automated numerous tasks such as the detection of AOVs, LPEs and Layers, and its deployment on the farm so that, in the end, all we have to do is select a folder containing images and the whole process is completed. 

[A] Given the high technical demands of our productions, in particular large quantities of Grooming, Scattering and special effects, the power and quality of the denoise algorithms we were using were insufficient. In addition, in order to meet the deadlines and quotas between the different productions on campus, we use a fairly low sampling rate to save as much time as possible on the renderfarm. All this led us to look for an alternative solution and, by chance, at the same time Renderman released its new AI-based denoiser.

Do you have any ambitions for this tool?

[T] We initially developed this tool specifically to meet our production needs, and then worked on making it more flexible so that it could be used more widely, firstly by other film groups on campus, and then more generally by anyone who wanted to use it.

[A] We would like to add Deadline support to cover the two main dispatchers on the market.

Where do you see yourself in 5 years’ time ? 

[A] In five year’s time, I would like to continue my learning in the field of special effects and tool development. I’m passionate about the idea of taking on new technical challenges and honing my skills in these constantly evolving fields, drawing in particular on my experience in creating Dopoise, the USD pipeline on EL DODORADO and the various tools I have designed. At the same time, I want to put my knowledge to good use by taking part in various productions. My aim is to contribute to many new films and series, bringing my knowledge to bear on increasingly complex technical challenges. 

[T] For me, I would like to work as a technical director for a renowned film studio and take part in film creations that, one day perhaps, will mark their era. I would also like to explore the field of research and development (R&D) so that I can collaborate and contribute to the development of technologies and production methods in our industry. 

Tom and Adrien have posted the link to download the tool on their respective websites : https://www.tom-perony.fr/ et https://www.lhabitantadrien.com/