en
fr

Expert's words: Philippe Llerena

TD lighting and compositing

10 Jan. 2020
Fabienne Legall Marketing Director
Fabienne Legall

Philippe Llerena

After graduating in science, Philippe Llerena chooses to use his knowledge in quantum physics and maths in the 3D field. He started his career as a free-lance compositing artist for prestigious studios (MPC, Unit Image, Mikros, Partizan…). As a lighting, compositing and rendering expert, he then worked at The Yard, On Animation or even Isart Digital School as a teacher, and also Khepris France as manager. He now is a Guerilla specialist at Mercenaries Engineering and Developer at Superprod. 

Playmobil The Movie 2019 - 2.9 Film Holding - Morgen production GmbH

Playmobil The Movie 2019 - 2.9 Film Holding - Morgen production GmbH

You have had the opportunity to work on major productions like Harry Potter, Minuscule or even Playmobil but also smaller and more personal projects. Does it affect the way you work on them? Is pressure commensurate with the size of the production?  
All these projects are not the same size. Harry Potter and Playmobil are industrial productions whereas Minuscule 2 for instance is close to crafted production. As a TD, the difference is not that much about pressure but about rigor and feasibility within the tools. For Playmobil for instance, the goal was to light, render and composit 1,500 shots in 3 months with a small team of graphic designers, some being beginers who did not need to know any subtleties of the pipeline but were asked to focus on their images. Communication is also one of the main differences and depends very much on the size of the teams. There is no room for unclear meessages without a trace. You can't do without a production monitoring tool. You exchange information fast without seeing your interlocutor and even without knowing if he or she is in India or downstairs. Supervisors probably don't manage the pressure the same way ,but it has never been my case.

You have a degree in mechanics and electromagnetism, you master hardware and rendering, you are a Python developer... and an artist. Do you think that art and science feed on each other? 
I have been nourrished by a quote from Robert A. Heinlein after my first reading of “Time Enough for Love”: ‘ A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects.' I would love to be this projection but unfortunately I don't manage to link science and tart and the other way round yet. It's probably why I feel comfortable in the VFX field as it requires less creation and more observation. I'm espacially talking about invisible effects like restore or set extensions. I am not very good at drawing and, for the sake of humanity, I only sing alone in my car. I have the feeling I hide a bit too much behind technique not to give my artistic opinion, but I am working on it. Not very far from us, I love Maa's work as I consider it is a great mix between art and technique. I assume both are so complex now that a 2020 Leonardo Da Vinci could not master one and the other in a single life. 

Minuscule: Mandibles from far away - MMXVIII Futurikon Films - IFilmfilm - France 3 Cinema

Minuscule: Mandibles from far away - MMXVIII Futurikon Films - IFilmfilm - France 3 Cinema


For the past 15 years, you have seen new technologies come and impact images creation and rendering. What do you think the major next steps will be? 

I am not a very good prophet even though I sometimes look like one (hair and beard) but I am very impressed by AI algorythms, whether they are denoisers or even new 2D rotoscopy algorythms that I saw at Siggraph last year. Incidentally, at Mercenaries Engineering, we noticed a bug when testing the Intel 0.9 denoiser: it recreated a Global Illumination when it was missing in the AOV. It leaves room for future GI algorythms in AI. I think it is just the beginning of creation assistance and that will lead to more complexity with less clicks. I also think procedural data generation is a major stake in the content creation . But I am also convinced that, before all that, there is a kind of wake-up call of small studios about industrial processes. I hear more and more talking about "pipeline" and I see with pride that my students tackle the subject of scalability. I think that the gap between the smallest and the biggest studios reduces. Mastering image making is more important than the technical gap. I was about to forget: USD, in short term. I start to have a look at it. Technology is not really revolutionnary, I even have the feeling that that it is less effective than other existing solutions but it has the advantage of standardizing a lot of things and offering more accessible assembling solutions. And once again let the possibility for smaller studios to use tools that are similar to the ones used by the biggest studios.