Animaj accelerates 3D pipeline to streamline kids’ animation with AWS

In less than three years, French startup Animaj made a name for itself in media production for kids. The company, founded in 2022, specializes in transforming intellectual property (IP) into bite-sized programming for 2-9 year-olds around the world. Animaj taps 3D animation vendors to deliver content for Pocoyo, Kidibli, Hey Kids, and Rabbids Invasion. It also maintains a robust internal pipeline built with Amazon Web Services (AWS) to accelerate production and create content more efficiently.

Animaj Vice President of Engineering Antoine Lhermitte, who oversees the company’s infrastructure team, explained, “AWS offers services and solutions tailored for the media industry, unlike other cloud providers, and they’re easy to deploy and scale so we can test new ideas quickly,” he shared. “We’ve found Amazon SageMaker and AWS Deadline Cloud to be very efficient; they make it easy to scale and manage new workflows from our environments.”

Supercharging creativity with artificial intelligence (AI)

Currently, Animaj comprises about 60 team members spread across offices in Paris, London, and Madrid. This includes engineering, general administration, production, consumer products, content, and communications teams to oversee vendor relationships, licensing, distribution, and promotion for each IP.

A typical season for one of Animaj’s animated programs comprises 52 episodes, each ranging anywhere from 3-8 minutes. Its programming runs on various platforms, including Netflix, YouTube, and local television channels. Every episode is created in house under the guidance of an Animaj creative director. Outside vendors are then enlisted for storyboarding, animatics, and 3D animation. Looking to expedite the animation process, Animaj began experimenting with motion capture (mocap) and Amazon SageMaker earlier this year. The company uses Move.AI to capture key poses for digital characters, then processes the data to return an output that matches the desired animated performance. To achieve this, the Animaj team developed a system using models trained in Amazon SageMaker.

These custom models automatically retarget human mocap data to its characters, reducing the work animators must do by hand so they can finish episodes more quickly. Despite historical time and budget constraints, the team found that mocap can be a good fit for characters that have proportions different from humans—like those in children’s cartoons.

For its latest production, Animaj retargeted mocap data to character models in Autodesk Maya 3D animation software but found an additional use for the trained models—accelerating the creation of animatics. Using early sketches from the directors, the trained models were able to predict subsequent character poses and translate them into 3D space. With sketches processed in minutes, animators could then import the files into Maya for refinement.

Spirit of experimentation

The team then turned to AWS Deadline Cloud for processing the scenes from Maya. Animaj’s engineers had been managing the software’s rendering fleets manually, but server set up and configuration work took away time the team could better spend innovating on its AI-based creative tools.

Leveraging AWS Deadline Cloud, Animaj was able to streamline its workflow. AWS Deadline Cloud automatically provisions the Maya environments, allowing engineers to scale render capacity up or down based on dynamic workloads. Animaj can adjust capacity in minutes rather than hours, accelerating animators’ creative feedback loops.

Integrating AWS Deadline Cloud also enabled Animaj to build a scalable backend for its machine learning pipeline. Data scientists preprocess mocap data for training on Amazon SageMaker. Then, the studio’s proprietary AI model generates scene layout suggestions that artists review in Maya. By offloading provisioning and scaling duties to AWS Deadline Cloud, the Animaj team is able to focus efforts on refining the AI assistant.

Rethinking 3D production

“The economics of a traditional 3D production pipeline are almost impossible, but with AWS and AI, it’s far more manageable. We can produce faster, iterate more, and expand our catalog,” Lhermitte shared. “We save money and time that we can reinvest in making shows that are more inspiring and better quality and create more characters that have added depth.”

In addition to Amazon SageMaker and AWS Deadline Cloud, Animaj uses Amazon Elastic Compute Cloud (Amazon EC2) and Amazon Simple Storage Service (Amazon S3) to support compute processing and storage needs.

Though AI has become an integral part of Animaj’s production process, Lhermitte emphasizes that AI is still just a tool. He concluded, “The quality of a show is going to depend on the creative power of the people working on it. We use AI to automate some of the more laborious workflow aspects, but still leave the most valuable contributions to the quality of our content to our team of expert artists.”

Learn more about how AWS services and solutions help creative studios produce, deliver, and monetize content, or get in touch with an AWS for Media & Entertainment representative.