Paul Debevec

Chloe LeGendre, Lukas Lepicovsky, Paul Debevec

While the LED panels used in virtual production systems can display vibrant imagery with a wide color gamut, they produce problematic color shifts when used as lighting due to their peaky spectral output from narrow-band red, green, and blue LEDs. In this work, we present an improved color calibration process for virtual production stages which ameliorates this color rendition problem while also passing through accurate in-camera background colors. We do this by optimizing linear color correction transformations for 1) the LED panel pixels visible in the field of view of the camera, 2) the pixels outside the field of view of the camera illuminating the subjects, and, as a post-process, 3) the pixel values recorded by the camera. The result is that footage shot in an RGB LED panel virtual production stage can exhibit more accurate skin tones and costume colors while still reproducing the desired colors of the in-camera background.

Ryan Burgert, Yuancheng Xu, Wenqi Xian et al.

Generative modeling aims to transform random noise into structured outputs. In this work, we enhance video diffusion models by allowing motion control via structured latent noise sampling. This is achieved by just a change in data: we pre-process training videos to yield structured noise. Consequently, our method is agnostic to diffusion model design, requiring no changes to model architectures or training pipelines. Specifically, we propose a novel noise warping algorithm, fast enough to run in real time, that replaces random temporal Gaussianity with correlated warped noise derived from optical flow fields, while preserving the spatial Gaussianity. The efficiency of our algorithm enables us to fine-tune modern video diffusion base models using warped noise with minimal overhead, and provide a one-stop solution for a wide range of user-friendly motion control: local object motion control, global camera movement control, and motion transfer. The harmonization between temporal coherence and spatial Gaussianity in our warped noise leads to effective motion control while maintaining per-frame pixel quality. Extensive experiments and user studies demonstrate the advantages of our method, making it a robust and scalable approach for controlling motion in video diffusion models. Video results are available on our webpage: https://eyeline-labs.github.io/Go-with-the-Flow. Source code and model checkpoints are available on GitHub: https://github.com/Eyeline-Labs/Go-with-the-Flow.

Li Ma, Mingming He, Xueming Yu et al.

Being able to relight human performance is a fundamental task for post production and content creation. We present BodyReLux, a subject-specific video diffusion-based framework for relighting full-body human performances in a temporally consistent way. Our model is trained on a hybrid dataset of pixel-aligned video relighting pairs, covering a diverse combination of lighting conditions, performances and viewpoints. To acquire such dataset, we combine traditional static One-Light-at-a-Time (OLAT) capture and a novel dynamic performance capture in which two smoothly varying lighting sequences are rapidly interleaved. Because the lighting operates above the human flicker-fusion threshold, the interleaving does not appear to strobe. We train our video relighting model from a pretrained text-to-video model to fully leverage the generative priors for producing high quality videos. To achieve accurate lighting control, we introduce a new lighting conditioning method that represents each light source as a token. We further condition on sequences of lighting using masked attention to support dynamic lighting control. Together with a carefully designed data augmentation pipeline, we achieve photorealistic, robust, and temporally consistent video relighting of subject-specific human performances.

Yimu Wang, Xuye Liu, Wei Pang et al.

Recent advances in diffusion models have revolutionized video generation, offering superior temporal consistency and visual quality compared to traditional generative adversarial networks-based approaches. While this emerging field shows tremendous promise in applications, it faces significant challenges in motion consistency, computational efficiency, and ethical considerations. This survey provides a comprehensive review of diffusion-based video generation, examining its evolution, technical foundations, and practical applications. We present a systematic taxonomy of current methodologies, analyze architectural innovations and optimization strategies, and investigate applications across low-level vision tasks such as denoising and super-resolution. Additionally, we explore the synergies between diffusionbased video generation and related domains, including video representation learning, question answering, and retrieval. Compared to the existing surveys (Lei et al., 2024a;b; Melnik et al., 2024; Cao et al., 2023; Xing et al., 2024c) which focus on specific aspects of video generation, such as human video synthesis (Lei et al., 2024a) or long-form content generation (Lei et al., 2024b), our work provides a broader, more updated, and more fine-grained perspective on diffusion-based approaches with a special section for evaluation metrics, industry solutions, and training engineering techniques in video generation. This survey serves as a foundational resource for researchers and practitioners working at the intersection of diffusion models and video generation, providing insights into both the theoretical frameworks and practical implementations that drive this rapidly evolving field. A structured list of related works involved in this survey is also available on https://github.com/Eyeline-Research/Survey-Video-Diffusion.

Gene Chou, Wenqi Xian, Guandao Yang et al. · deepmind

A versatile video depth estimation model should (1) be accurate and consistent across frames, (2) produce high-resolution depth maps, and (3) support real-time streaming. We propose FlashDepth, a method that satisfies all three requirements, performing depth estimation on a 2044x1148 streaming video at 24 FPS. We show that, with careful modifications to pretrained single-image depth models, these capabilities are enabled with relatively little data and training. We evaluate our approach across multiple unseen datasets against state-of-the-art depth models, and find that ours outperforms them in terms of boundary sharpness and speed by a significant margin, while maintaining competitive accuracy. We hope our model will enable various applications that require high-resolution depth, such as video editing, and online decision-making, such as robotics. We release all code and model weights at https://github.com/Eyeline-Research/FlashDepth

Christophe Bolduc, Julien Philip, Li Ma et al.

We present Lighting in Motion (LiMo), a diffusion-based approach to spatiotemporal lighting estimation. LiMo targets both realistic high-frequency detail prediction and accurate illuminance estimation. To account for both, we propose generating a set of mirrored and diffuse spheres at different exposures, based on their 3D positions in the input. Making use of diffusion priors, we fine-tune powerful existing diffusion models on a large-scale customized dataset of indoor and outdoor scenes, paired with spatiotemporal light probes. For accurate spatial conditioning, we demonstrate that depth alone is insufficient and we introduce a new geometric condition to provide the relative position of the scene to the target 3D position. Finally, we combine diffuse and mirror predictions at different exposures into a single HDRI map leveraging differentiable rendering. We thoroughly evaluate our method and design choices to establish LiMo as state-of-the-art for both spatial control and prediction accuracy.

Yuancheng Xu, Wenqi Xian, Li Ma et al.

We introduce a framework that enables both multi-view character consistency and 3D camera control in video diffusion models through a novel customization data pipeline. We train the character consistency component with recorded volumetric capture performances re-rendered with diverse camera trajectories via 4D Gaussian Splatting (4DGS), lighting variability obtained with a video relighting model. We fine-tune state-of-the-art open-source video diffusion models on this data to provide strong multi-view identity preservation, precise camera control, and lighting adaptability. Our framework also supports core capabilities for virtual production, including multi-subject generation using two approaches: joint training and noise blending, the latter enabling efficient composition of independently customized models at inference time; it also achieves scene and real-life video customization as well as control over motion and spatial layout during customization. Extensive experiments show improved video quality, higher personalization accuracy, and enhanced camera control and lighting adaptability, advancing the integration of video generation into virtual production. Our project page is available at: https://eyeline-labs.github.io/Virtually-Being.

Haonan Qiu, Ning Yu, Ziqi Huang et al.

Visual diffusion models achieve remarkable progress, yet they are typically trained at limited resolutions due to the lack of high-resolution data and constrained computation resources, hampering their ability to generate high-fidelity images or videos at higher resolutions. Recent efforts have explored tuning-free strategies to exhibit the untapped potential higher-resolution visual generation of pre-trained models. However, these methods are still prone to producing low-quality visual content with repetitive patterns. The key obstacle lies in the inevitable increase in high-frequency information when the model generates visual content exceeding its training resolution, leading to undesirable repetitive patterns deriving from the accumulated errors. In this work, we propose CineScale, a novel inference paradigm to enable higher-resolution visual generation. To tackle the various issues introduced by the two types of video generation architectures, we propose dedicated variants tailored to each. Unlike existing baseline methods that are confined to high-resolution T2I and T2V generation, CineScale broadens the scope by enabling high-resolution I2V and V2V synthesis, built atop state-of-the-art open-source video generation frameworks. Extensive experiments validate the superiority of our paradigm in extending the capabilities of higher-resolution visual generation for both image and video models. Remarkably, our approach enables 8k image generation without any fine-tuning, and achieves 4k video generation with only minimal LoRA fine-tuning. Generated video samples are available at our website: https://eyeline-labs.github.io/CineScale/.

Kuan Heng Lin, Zhizheng Liu, Pablo Salamanca et al.

We present Vista4D, a robust and flexible video reshooting framework that grounds the input video and target cameras in a 4D point cloud. Specifically, given an input video, our method re-synthesizes the scene with the same dynamics from a different camera trajectory and viewpoint. Existing video reshooting methods often struggle with depth estimation artifacts of real-world dynamic videos, while also failing to preserve content appearance and failing to maintain precise camera control for challenging new trajectories. We build a 4D-grounded point cloud representation with static pixel segmentation and 4D reconstruction to explicitly preserve seen content and provide rich camera signals, and we train with reconstructed multiview dynamic data for robustness against point cloud artifacts during real-world inference. Our results demonstrate improved 4D consistency, camera control, and visual quality compared to state-of-the-art baselines under a variety of videos and camera paths. Moreover, our method generalizes to real-world applications such as dynamic scene expansion and 4D scene recomposition. See our project page for results, code, and models: https://eyeline-labs.github.io/Vista4D

Yiqun Mei, Mingming He, Li Ma et al.

Video portrait relighting remains challenging because the results need to be both photorealistic and temporally stable. This typically requires a strong model design that can capture complex facial reflections as well as intensive training on a high-quality paired video dataset, such as dynamic one-light-at-a-time (OLAT). In this work, we introduce Lux Post Facto, a novel portrait video relighting method that produces both photorealistic and temporally consistent lighting effects. From the model side, we design a new conditional video diffusion model built upon state-of-the-art pre-trained video diffusion model, alongside a new lighting injection mechanism to enable precise control. This way we leverage strong spatial and temporal generative capability to generate plausible solutions to the ill-posed relighting problem. Our technique uses a hybrid dataset consisting of static expression OLAT data and in-the-wild portrait performance videos to jointly learn relighting and temporal modeling. This avoids the need to acquire paired video data in different lighting conditions. Our extensive experiments show that our model produces state-of-the-art results both in terms of photorealism and temporal consistency.

Zhengming Yu, Li Ma, Mingming He et al.

Most digital videos are stored in 8-bit low dynamic range (LDR) formats, where much of the original high dynamic range (HDR) scene radiance is lost due to saturation and quantization. This loss of highlight and shadow detail precludes mapping accurate luminance to HDR displays and limits meaningful re-exposure in post-production workflows. Although techniques have been proposed to convert LDR images to HDR through dynamic range expansion, they struggle to restore realistic detail in the over- and underexposed regions. To address this, we present DiffHDR, a framework that formulates LDR-to-HDR conversion as a generative radiance inpainting task within the latent space of a video diffusion model. By operating in Log-Gamma color space, DiffHDR leverages spatio-temporal generative priors from a pretrained video diffusion model to synthesize plausible HDR radiance in over- and underexposed regions while recovering the continuous scene radiance of the quantized pixels. Our framework further enables controllable LDR-to-HDR video conversion guided by text prompts or reference images. To address the scarcity of paired HDR video data, we develop a pipeline that synthesizes high-quality HDR video training data from static HDRI maps. Extensive experiments demonstrate that DiffHDR significantly outperforms state-of-the-art approaches in radiance fidelity and temporal stability, producing realistic HDR videos with considerable latitude for re-exposure.

Youming Deng, Wenqi Xian, Guandao Yang et al.

In this paper, we present a self-calibrating framework that jointly optimizes camera parameters, lens distortion and 3D Gaussian representations, enabling accurate and efficient scene reconstruction. In particular, our technique enables high-quality scene reconstruction from Large field-of-view (FOV) imagery taken with wide-angle lenses, allowing the scene to be modeled from a smaller number of images. Our approach introduces a novel method for modeling complex lens distortions using a hybrid network that combines invertible residual networks with explicit grids. This design effectively regularizes the optimization process, achieving greater accuracy than conventional camera models. Additionally, we propose a cubemap-based resampling strategy to support large FOV images without sacrificing resolution or introducing distortion artifacts. Our method is compatible with the fast rasterization of Gaussian Splatting, adaptable to a wide variety of camera lens distortion, and demonstrates state-of-the-art performance on both synthetic and real-world datasets.

Ziqi Huang, Ning Yu, Gordon Chen et al.

Recent video generation models can produce smooth and visually appealing clips, but they often struggle to synthesize complex dynamics with a coherent chain of consequences. Accurately modeling visual outcomes and state transitions over time remains a core challenge. In contrast, large language and multimodal models (e.g., GPT-4o) exhibit strong visual state reasoning and future prediction capabilities. To bridge these strengths, we introduce VChain, a novel inference-time chain-of-visual-thought framework that injects visual reasoning signals from multimodal models into video generation. Specifically, VChain contains a dedicated pipeline that leverages large multimodal models to generate a sparse set of critical keyframes as snapshots, which are then used to guide the sparse inference-time tuning of a pre-trained video generator only at these key moments. Our approach is tuning-efficient, introduces minimal overhead and avoids dense supervision. Extensive experiments on complex, multi-step scenarios show that VChain significantly enhances the quality of generated videos.

Qian Wang, Ziqi Huang, Ruoxi Jia et al.

Despite recent advances, long-sequence video generation frameworks still suffer from significant limitations: poor assistive capability, suboptimal visual quality, and limited expressiveness. To mitigate these limitations, we propose MAViS, a multi-agent collaborative framework designed to assist in long-sequence video storytelling by efficiently translating ideas into visual narratives. MAViS orchestrates specialized agents across multiple stages, including script writing, shot designing, character modeling, keyframe generation, video animation, and audio generation. In each stage, agents operate under the 3E Principle -- Explore, Examine, and Enhance -- to ensure the completeness of intermediate outputs. Considering the capability limitations of current generative models, we propose the Script Writing Guidelines to optimize compatibility between scripts and generative tools. Experimental results demonstrate that MAViS achieves state-of-the-art performance in assistive capability, visual quality, and video expressiveness. Its modular framework further enables scalability with diverse generative models and tools. With just a brief idea description, MAViS enables users to rapidly explore diverse visual storytelling and creative directions for sequential video generation by efficiently producing high-quality, complete long-sequence videos. To the best of our knowledge, MAViS is the only framework that provides multimodal design output -- videos with narratives and background music.

Pascal Clausen, Li Ma, Mingming He et al.

We present a technique for fitting high dynamic range illumination (HDRI) sequences using anisotropic spherical Gaussians (ASGs) while preserving temporal consistency in the compressed HDRI maps. Our approach begins with an optimization network that iteratively minimizes a composite loss function, which includes both reconstruction and diffuse losses. This allows us to represent all-frequency signals with a small number of ASGs, optimizing their directions, sharpness, and intensity simultaneously for an individual HDRI. To extend this optimization into the temporal domain, we introduce a temporal consistency loss, ensuring a consistent approximation across the entire HDRI sequence.

Xiuming Zhang, Pratul P. Srinivasan, Boyang Deng et al.

We address the problem of recovering the shape and spatially-varying reflectance of an object from multi-view images (and their camera poses) of an object illuminated by one unknown lighting condition. This enables the rendering of novel views of the object under arbitrary environment lighting and editing of the object's material properties. The key to our approach, which we call Neural Radiance Factorization (NeRFactor), is to distill the volumetric geometry of a Neural Radiance Field (NeRF) [Mildenhall et al. 2020] representation of the object into a surface representation and then jointly refine the geometry while solving for the spatially-varying reflectance and environment lighting. Specifically, NeRFactor recovers 3D neural fields of surface normals, light visibility, albedo, and Bidirectional Reflectance Distribution Functions (BRDFs) without any supervision, using only a re-rendering loss, simple smoothness priors, and a data-driven BRDF prior learned from real-world BRDF measurements. By explicitly modeling light visibility, NeRFactor is able to separate shadows from albedo and synthesize realistic soft or hard shadows under arbitrary lighting conditions. NeRFactor is able to recover convincing 3D models for free-viewpoint relighting in this challenging and underconstrained capture setup for both synthetic and real scenes. Qualitative and quantitative experiments show that NeRFactor outperforms classic and deep learning-based state of the art across various tasks. Our videos, code, and data are available at people.csail.mit.edu/xiuming/projects/nerfactor/.

Loc Huynh, Bipin Kishore, Paul Debevec

We present a learning-based method for estimating 4D reflectance field of a person given video footage illuminated under a flat-lit environment of the same subject. For training data, we use one light at a time to illuminate the subject and capture the reflectance field data in a variety of poses and viewpoints. We estimate the lighting environment of the input video footage and use the subject's reflectance field to create synthetic images of the subject illuminated by the input lighting environment. We then train a deep convolutional neural network to regress the reflectance field from the synthetic images. We also use a differentiable renderer to provide feedback for the network by matching the relit images with the input video frames. This semi-supervised training scheme allows the neural network to handle unseen poses in the dataset as well as compensate for the lighting estimation error. We evaluate our method on the video footage of the real Holocaust survivors and show that our method outperforms the state-of-the-art methods in both realism and speed.

Peter Hedman, Pratul P. Srinivasan, Ben Mildenhall et al.

Neural volumetric representations such as Neural Radiance Fields (NeRF) have emerged as a compelling technique for learning to represent 3D scenes from images with the goal of rendering photorealistic images of the scene from unobserved viewpoints. However, NeRF's computational requirements are prohibitive for real-time applications: rendering views from a trained NeRF requires querying a multilayer perceptron (MLP) hundreds of times per ray. We present a method to train a NeRF, then precompute and store (i.e. "bake") it as a novel representation called a Sparse Neural Radiance Grid (SNeRG) that enables real-time rendering on commodity hardware. To achieve this, we introduce 1) a reformulation of NeRF's architecture, and 2) a sparse voxel grid representation with learned feature vectors. The resulting scene representation retains NeRF's ability to render fine geometric details and view-dependent appearance, is compact (averaging less than 90 MB per scene), and can be rendered in real-time (higher than 30 frames per second on a laptop GPU). Actual screen captures are shown in our video.

Tiancheng Sun, Zexiang Xu, Xiuming Zhang et al.

The light stage has been widely used in computer graphics for the past two decades, primarily to enable the relighting of human faces. By capturing the appearance of the human subject under different light sources, one obtains the light transport matrix of that subject, which enables image-based relighting in novel environments. However, due to the finite number of lights in the stage, the light transport matrix only represents a sparse sampling on the entire sphere. As a consequence, relighting the subject with a point light or a directional source that does not coincide exactly with one of the lights in the stage requires interpolation and resampling the images corresponding to nearby lights, and this leads to ghosting shadows, aliased specularities, and other artifacts. To ameliorate these artifacts and produce better results under arbitrary high-frequency lighting, this paper proposes a learning-based solution for the "super-resolution" of scans of human faces taken from a light stage. Given an arbitrary "query" light direction, our method aggregates the captured images corresponding to neighboring lights in the stage, and uses a neural network to synthesize a rendering of the face that appears to be illuminated by a "virtual" light source at the query location. This neural network must circumvent the inherent aliasing and regularity of the light stage data that was used for training, which we accomplish through the use of regularized traditional interpolation methods within our network. Our learned model is able to produce renderings for arbitrary light directions that exhibit realistic shadows and specular highlights, and is able to generalize across a wide variety of subjects.

Xiuming Zhang, Sean Fanello, Yun-Ta Tsai et al.

The light transport (LT) of a scene describes how it appears under different lighting and viewing directions, and complete knowledge of a scene's LT enables the synthesis of novel views under arbitrary lighting. In this paper, we focus on image-based LT acquisition, primarily for human bodies within a light stage setup. We propose a semi-parametric approach to learn a neural representation of LT that is embedded in the space of a texture atlas of known geometric properties, and model all non-diffuse and global LT as residuals added to a physically-accurate diffuse base rendering. In particular, we show how to fuse previously seen observations of illuminants and views to synthesize a new image of the same scene under a desired lighting condition from a chosen viewpoint. This strategy allows the network to learn complex material effects (such as subsurface scattering) and global illumination, while guaranteeing the physical correctness of the diffuse LT (such as hard shadows). With this learned LT, one can relight the scene photorealistically with a directional light or an HDRI map, synthesize novel views with view-dependent effects, or do both simultaneously, all in a unified framework using a set of sparse, previously seen observations. Qualitative and quantitative experiments demonstrate that our neural LT (NLT) outperforms state-of-the-art solutions for relighting and view synthesis, without separate treatment for both problems that prior work requires.

Chloe LeGendre, Wan-Chun Ma, Rohit Pandey et al.

We present a learning-based technique for estimating high dynamic range (HDR), omnidirectional illumination from a single low dynamic range (LDR) portrait image captured under arbitrary indoor or outdoor lighting conditions. We train our model using portrait photos paired with their ground truth environmental illumination. We generate a rich set of such photos by using a light stage to record the reflectance field and alpha matte of 70 diverse subjects in various expressions. We then relight the subjects using image-based relighting with a database of one million HDR lighting environments, compositing the relit subjects onto paired high-resolution background imagery recorded during the lighting acquisition. We train the lighting estimation model using rendering-based loss functions and add a multi-scale adversarial loss to estimate plausible high frequency lighting detail. We show that our technique outperforms the state-of-the-art technique for portrait-based lighting estimation, and we also show that our method reliably handles the inherent ambiguity between overall lighting strength and surface albedo, recovering a similar scale of illumination for subjects with diverse skin tones. We demonstrate that our method allows virtual objects and digital characters to be added to a portrait photograph with consistent illumination. Our lighting inference runs in real-time on a smartphone, enabling realistic rendering and compositing of virtual objects into live video for augmented reality applications.

John Flynn, Michael Broxton, Paul Debevec et al.

We present a novel approach to view synthesis using multiplane images (MPIs). Building on recent advances in learned gradient descent, our algorithm generates an MPI from a set of sparse camera viewpoints. The resulting method incorporates occlusion reasoning, improving performance on challenging scene features such as object boundaries, lighting reflections, thin structures, and scenes with high depth complexity. We show that our method achieves high-quality, state-of-the-art results on two datasets: the Kalantari light field dataset, and a new camera array dataset, Spaces, which we make publicly available.

Tiancheng Sun, Jonathan T. Barron, Yun-Ta Tsai et al.

Lighting plays a central role in conveying the essence and depth of the subject in a portrait photograph. Professional photographers will carefully control the lighting in their studio to manipulate the appearance of their subject, while consumer photographers are usually constrained to the illumination of their environment. Though prior works have explored techniques for relighting an image, their utility is usually limited due to requirements of specialized hardware, multiple images of the subject under controlled or known illuminations, or accurate models of geometry and reflectance. To this end, we present a system for portrait relighting: a neural network that takes as input a single RGB image of a portrait taken with a standard cellphone camera in an unconstrained environment, and from that image produces a relit image of that subject as though it were illuminated according to any provided environment map. Our method is trained on a small database of 18 individuals captured under different directional light sources in a controlled light stage setup consisting of a densely sampled sphere of lights. Our proposed technique produces quantitatively superior results on our dataset's validation set compared to prior works, and produces convincing qualitative relighting results on a dataset of hundreds of real-world cellphone portraits. Because our technique can produce a 640 $\times$ 640 image in only 160 milliseconds, it may enable interactive user-facing photographic applications in the future.

Chloe LeGendre, Wan-Chun Ma, Graham Fyffe et al.

We present a learning-based method to infer plausible high dynamic range (HDR), omnidirectional illumination given an unconstrained, low dynamic range (LDR) image from a mobile phone camera with a limited field of view (FOV). For training data, we collect videos of various reflective spheres placed within the camera's FOV, leaving most of the background unoccluded, leveraging that materials with diverse reflectance functions reveal different lighting cues in a single exposure. We train a deep neural network to regress from the LDR background image to HDR lighting by matching the LDR ground truth sphere images to those rendered with the predicted illumination using image-based relighting, which is differentiable. Our inference runs at interactive frame rates on a mobile device, enabling realistic rendering of virtual objects into real scenes for mobile mixed reality. Training on automatically exposed and white-balanced videos, we improve the realism of rendered objects compared to the state-of-the art methods for both indoor and outdoor scenes.