How do you transform authentic archival footage into a seamless work of fiction?

Selected for Cannes Classics, The Golden Age (L'Âge d'Or) challenged conventional visual effects by blending documentary archives with newly filmed performances, allowing its characters to exist naturally within real moments from history.

In this interview, director Bérenger Thouin and Hugues Namur, VFX Supervisor at The Mill (ex-MPC Paris), discuss the creative collaboration that began long before production. Together, they reveal how visual effects became part of the film's storytelling language, combining traditional compositing, face replacement techniques, deep learning workflows and meticulous practical filmmaking to create an immersive cinematic experience.

Discover how archival footage was transformed into living cinema through months of creative experimentation, precise performances by Souheila Yacoub, and innovative visual effects designed to make audiences forget where documentary ends and fiction begins.

???? The Golden Age (L'Âge d'Or)
Selected for Cannes Classics

Featuring:
• Bérenger Thouin, Director
• Hugues Namur, VFX Supervisor, The Mill

ABOUT US
The Mill is a world-leading creative studio built on nearly four decades of award-winning craft. Spanning Brand & Content and Film & Series, we create visually ambitious work for the world's most discerning brands, studios and storytellers — conceived from the outset to move across platforms, markets and audiences without compromise.

FOLLOW US
Website: https://www.themill.com/
LinkedIn: https://www.linkedin.com/company/the-mill
Instagram: https://www.instagram.com/millchannel
X: https://x.com/MILL__CHANNEL

#TheMill #VFX #VFXBreakdown #CannesClassics #Filmmaking #VisualEffects #PostProduction #AI #DeepLearning #Cinema #BehindTheScenes

Here's a look at how Epic's artists approach concept work. From blank canvas to final concepts, they sketch, block out, and refine using traditional tools and evolving ones like Nano Banana, with the artist’s vision driving the ever-evolving process from start to finish.

Conventional physically based rendering (PBR) pipelines generate photorealistic images through computationally intensive light transport simulations. Although recent deep learning approaches leverage diffusion model priors with geometry buffers (G-buffers) to produce visually compelling results without explicit scene geometry or light simulation, they remain constrained by two major limitations. First, the iterative nature of the diffusion process introduces substantial latency. Second, the inherent stochasticity of these generative models compromises physical accuracy and temporal consistency. In response to these challenges, we propose a novel, end-to-end, deterministic, single-step neural rendering framework, RenderFlow, built upon a flow matching paradigm. To further strengthen both rendering quality and generalization, we propose an efficient and effective module for sparse keyframe guidance. Our method significantly accelerates the rendering process and, by optionally incorporating sparsely rendered keyframes as guidance, enhances both the physical plausibility and overall visual quality of the output. The resulting pipeline achieves near real-time performance with photorealistic rendering quality, effectively bridging the gap between the efficiency of modern generative models and the precision of traditional physically based rendering. Furthermore, we demonstrate the versatility of our framework by introducing a lightweight, adapter-based module that efficiently repurposes the pretrained forward model for the inverse rendering task of intrinsic decomposition.
Publication Link: https://studios.disneyresearch.com/2026/05/31/renderflow-single-step-neural-rendering-via-flow-matching/

Soft boundaries, like thin hairs, are commonly observed in
natural and computer-generated imagery, but they remain
challenging for 3D vision due to the ambiguous mixing of
foreground and background cues. This paper introduces
Guardians of the Hair (HairGuard), a framework designed
to recover fine-grained soft boundary details in 3D vision
tasks. Specifically, we first propose a novel data curation
pipeline that leverages image matting datasets for training
and design a depth fixer network to automatically identify
soft boundary regions. With a gated residual module, the
depth fixer refines depth precisely around soft boundaries
while maintaining global depth quality, allowing plug-andplay
integration with state-of-the-art depth models. For
view synthesis, we perform depth-based forward warping
to retain high-fidelity textures, followed by a generative
scene painter that fills disoccluded regions and eliminates
redundant background artifacts within soft boundaries. Finally,
a color fuser adaptively combines warped and inpainted
results to produce novel views with consistent geometry
and fine-grained details. Extensive experiments
demonstrate that HairGuard achieves state-of-the-art performance
across monocular depth estimation, stereo image/
video conversion, and novel view synthesis, with significant
improvements in soft boundary regions.

Publication Link: https://studios.disneyresearch.com/2026/05/31/guardians-of-the-hair-rescuing-soft-boundaries-in-depth-stereo-and-novel-views/

Un pequeño resumen de lo ultimo que hemos hecho

Facial animation is one of the most labor-intensive aspects of animation and VFX, as traditional rigging consumes weeks of expert time and forces animators to spend countless hours manipulating hundreds of controls to achieve varied expressions. This technical complexity creates a barrier between artistic vision and execution, limiting creative exploration and iteration. In this paper, we introduce CANRig, a fully automated neural facial rigging approach that simplifies the process of creating and editing facial poses by benefiting from global correlations learned from data. Unlike existing neural face models that either sacrifice local control or demand extensive manual region setup, our method introduces continuous local control through a novel conditioning mechanism that operates on a variable region. By modeling deformation as cross-attention between control handles and mesh vertices—modulated by a user-defined region—we enable seamless transitions from precise local adjustments to broad global changes. We further expand our method with a shape-preserving workflow that enables iterative edits, guaranteeing that changes remain untouched even as controls are reconfigured. Our method delivers the best of both worlds: the automation and naturalness of neural methods with the granular control that professional animators demand, and we demonstrate its effectiveness across multiple applications in both animation and high-end visual effects pipelines.

Publication Link: https://studios.disneyresearch.com/2026/04/04/canrig-cross-attention-neural-face-rigging-with-variable-local-control/

Human motion data is inherently rich and complex, containing both semantic content and subtle stylistic features that are challenging to model. We propose a novel method for effective disentanglement of the style and content in human motion data to facilitate style transfer. Our approach is guided by the insight that content corresponds to coarse motion attributes while style captures the finer, expressive details. To model this hierarchy, we employ Residual Vector Quantized Variational Autoencoders (RVQ-VAEs) to learn a coarse-to-fine representation of motion. We further enhance the disentanglement by integrating codebook learning with contrastive learning and a novel information leakage loss to organize the content and the style across different codebooks. We harness this disentangled representation using our simple and effective inference-time technique Quantized Code Swapping, which enables motion style transfer without requiring any fine-tuning for unseen styles. Our framework demonstrates strong versatility across multiple inference applications, including style transfer, style removal, and motion blending.

Link to publication: https://studios.disneyresearch.com/2026/04/04/vq-style-disentangling-style-and-content-in-motion-with-residual-quantized-representations/

You're watching a tech demo in action! Our Anvil engine uses Micropolygons to render millions of geometric details in real time. This demo showcases the scale of geometry the system can handle.

Learn more about Ubisoft: www.ubisoftgroup.com

© 2026 Ubisoft Entertainment. All Rights Reserved. Ubisoft and the Ubisoft logo are registered trademarks in the US and/or other countries.