This article is also available in:
Français
Another day at SIGGRAPH! Yet again, many interesting sessions here at Denver. Moreover, the exhibition floor is now open.
Cap Digital Reflects on NVIDIA and Meta’s Stance on AI
To kick off the day, we interviewed Marc Bourhis from Cap Digital, a French competitiveness cluster. He shared his thoughts on what NVIDIA CEO Jensen Huang and Meta CEO Mark Zuckerberg said on AI.
Although Bourhis felt somewhat unsatisfied with the lack of major announcements, he noted that the presentations confirmed certain trends and partnerships, particularly between Shutterstock and NVIDIA. He also discussed the focus on avatars and smart agents, which are central to both companies’ strategies. Additionally, he highlighted some gaps and limitations in what was said by those two tech giants, particularly regarding copyright issues and the ecological impact of AI.
The presentations from NVIDIA and Meta are available below.
Exhibition floor!
Another highlight of the day was the opening of the exhibition hall. As you may have seen in the Instagram video we shared, the attendees were eager to chat with the exhibitors. Even though SIGGRAPH 2024 doesn’t have a massive exhibition floor, people are enthusiastic.
SIGGRAPH 2024: France in the Spotlight
SIGGRAPH also features the Electronic Theater / Animation Theater, showcasing projects from around the world. Notably, French animated projects were prominently featured: 1/3 of the selected projects were created in France. These are mostly student films, indicating that top French animation schools continue to create highly-skilled graduates. Moreover, the proportion of French projects has been increasing over the years, as shown by the data compiled by Pierre Hénon.
Short films from schools such as MOPA, ENSI, ESMA, Supinfocom, Piktura, and ARTFX are featured this year. Notably, Après Papi, a short film from MoPA, was awarded the Best Student Project award. Congratulations to the team: Swann Valenza, Florian Gomes Freitas, Juliette Michel, Victoria Leviaux, and Axel Sence.
The selection also includes Au Revoir Mon Monde, a MoPA short film featured on 3DVF (read our interview with the team), Silhouette from ARTFX, and Vidalia from ESMA (watch the trailer).
Lips Don’t Lie: AI in Animation
In the afternoon, a conference on machine learning and AI for animation took place. Three companies showcased their innovations:
- Square Enix showcased Lip-Sync ML, a machine learning framework that generates lip-sync from audio recordings. The studio used clips and animations from previous games as datasets to create lip-sync for Final Fantasy VII Rebirth. Slides and the publication are available on the studio’s website.
- Golaem told us about Audio2Rig, another lip-sync tool. Known primarily for its crowd simulation solution, Golaem explained that Audio2Rig is a deep learning tool that uses previously animated sequences to generate facial animation and lip-sync for new sequences. It is compatible with any production rig. The prototype demo video is available below. Currently, Audio2Rig is an RD prototype and not a commercial product. Blue Spirit studio is involved in the project.
- Boxel Studio offered a less technical presentation. The team discussed their work on the Superman and Lois series, where they had to find new solutions to meet tight deadlines, utilizing AI in various aspects. This approach allowed the studio to animate Doomsday.
SIGGRAPH, Day 4
We will continue to cover the event in the coming days.
To stay updated on our content, be sure to follow us on social media: Facebook, X/Twitter, Instagram, LinkedIn, and YouTube.
