Short Film: To-Do

In the course of a year and a half, students from Switzerland created a short film. Their school recently switched to Blender and with a few exceptions, they managed to create the entire project within Blender. I talked to Samuel Bienz, one of the directors.

Our school is called “Hochschule Luzern” (HSLU) which is one of the Schools of Applied Science in central Switzerland and I studied 3 years for a Bachelor in Animation.

Our Bachelor course is very open to all sorts of styles and narratives. At the beginning of our studies everybody gets inputs in the different animation techniques 2D, Stop Motion and 3D. Later during personal or applied projects we can deepen the knowledge in the techniques we choose for ourselves. We also learned about storytelling, world-building, pipeline, animation history and sounddesign. (Our year was the first that got introduced to 3D with Blender and the experience was much better than when I tried Cinema4D years ago. At first I wanted to be a 2D animator but that course changed my mind.)

“To-Do” wasn’t always what it is today. At the start of our last year in about August 2021 every student had the chance to pitch a film idea which students and faculty voted on. I was overworked at the time to say the least but to my surprise my pitch actually went through and I was joined by my two partners in crime Iris and Pauline. Which were a pretty good fit as we all had different fields of expertise. Iris focused on animation but she also worked on Sounddesign, Editing and general Organization of the Project. Pauline focused on Modeling and Texturing but also worked on Concept Art and Design. I focused on filling the gaps between and worked on Story, Layout, Rigging, Light, Pipeline, Procedurals and Compositing. During our story development we realized that my original idea was too abstract and big to be realized in the time frame we had. We decided to re-write everything but keep the core message as intact as possible and thus “To-Do” was born. A story about a girl trying to organize her ever growing work using sticky notes and getting overwhelmed by this task. Among other things she symbolizes a work ethic that we did not intend to follow and we swore to ourselves to work against crunch as much as we could.

From the very start it was clear that we would use Blender as our main application. It would be where we make the storyboard, animatic, layout, animation, generative assets, shader, lighting and rendering as well as some simulations, compositing and editing of our film. Especially useful would be the then still pretty new Asset Browser and Linked Libraries, which I loved to use in my previous project. But not everything was optimal to be done in Blender for various reasons.

While Iris and me knew our ways around Blender pretty well already, Pauline, who came to us after a gap year, had more experience using ZBrush for modeling. Because of that and the urgency to get the character up and running as soon as possible, we decided to model the character in ZBrush and then send it to Blender for re-topology, texturing and rigging.

While we did all the texturing inside Blender it was easier to make and export custom brushes in Photoshop. The textures were then saved into our assetbrowser and added to the shader trees of our assets. These assets were then linked to the scenes. We really tried to get everything out of the Blender and make a workflow possible that allowed for all three of us to work on the project at the same time. With this workflow it was for the most part pretty easy to iterate on textures and assets. I say for the most part because this was before you could mark Collections as Assets and the workaround of using “Instance Collections” as assets could sometimes break or not load properly. Nothing that couldn’t be fixed manually though.

Most problems we encountered had to do with simulation and “cross platform compatibility” for a lack of better words. Blender was able to handle most of the simulations very well but there has been a couple cases, where we had no other choice than to switch to Houdini. While most scenes that have to do with sticky notes have been solved using Gemoetry Nodes and the help of the Blender community on social media, we have some transitions and effects in the film where hundreds of sticky notes are flying through the air and Blender simply could not handle simulating the physics of them. Simulating the notes in Houdini and then exporting them as an Alembic did the trick. Blender was still struggling a little bit because of the vertex count when importing but everything was fine in render time.

The goal with our style was to capture some sort of rushed and unfinished but nostalgic feeling - Imagine the YouTube “Lofi-Girl” but she’s stressed out. In addition to trying to achieve this feeling, my wish was to create something unique and keep the render times as low as possible. For months were were struggling to get a look that we liked until I had an epiphany. A very simple and very Blender specific idea. In contrast to the films story I asked myself, why should WE be doing the hard work? You see, while playing around with the sample count and Viewport Denoiser in Cycles, I noticed that in the first few seconds the image in the Viewport looked almost like a watercolor painting. But this would surely not look good as a moving image, right? The noise will make everything too erratic, right? Actually, no. It looked pretty much exactly like what we were searching for with minimal Compositing needed and best of all, the render times were only about 20 seconds per frame – dopamin for my impatient brain.

One of the greatest aspects of working with Blender on this project has been the community. When my team and I were stuck there has always been some person or resource that would lead to a solution and not only that but the active and rapid development of Blender would lead to a very fluid workflow where we would inform ourselves about newer features and could experiment and implement them if we wanted to.

As I said previously the pitch happened in late 2021. Preproduction ended some time in January 2022 and we have been working tirelessly but without crunch until mid June 2022 when we had to hand the project in for grading by the school. After a break of a couple of weeks and everyone passing their Bachelor degree we continued working on the project to fix some animations and render mistakes off and on for another half a year while working separately on our portfolios and looking for jobs. The progress was pretty slow because essentially at the end we only had one machine that would be able to handle rendering everything and wouldn’t break liked connections or crash due to the graphics card not being able to handle the hair particle system and that machine was located at the school while all of us were moving away from school. In December we were finally done with the visuals and only needed to do the final sound mix which happened in early January 2023. And like that “To-Do” has been done…

Well, not quite. Now that we’re done with the film, it’s time to start introducing our baby to the short film festivals. We’re currently working on the promotion of the film and are handing it in to several festivals all over the world. Fingers crossed you can see it somewhere soon!

There’s a lot I’ve learned in this project but maybe one of the biggest things to take away from the project that goes back to very meaning of the film is that you shouldn’t let work define your life and that you don’t need to “earn” breaks. You don’t need to work hard or achieve anything to take a break. You just simply deserve them.