cs.CVOct 22, 2026

SyncVid: World-Equivariant Diffusion for Synchronized Video Generation

Authors: Farhad G. Zanjani, Animesh Karnewar, Fatih Porikli, Amirhossein Habibian

Organizations: Qualcomm AI Research is an initiative of Qualcomm Technologies, Inc.

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Generating multiple videos of a dynamic scene from a single reference video requires synthesizing both observed and unobserved content consistently across viewpoints. Existing camera-control methods for video diffusion typically rely on architectural mechanisms, including camera embeddings, cross-view attention, and geometric conditioning. Although effective for static scene content, these approaches often struggle with the fine-grained, dynamic geometry that must remain coherent across views and over time. We synchronize the diffusion process itself by sampling the Gaussian noise prior in a canonical 3D space derived from the source video and rasterizing it into target cameras, rather than sampling independently in each 2D image plane. Our method, SyncVid, lifts a monocular video into a canonical 4D representation, associates noise with scene geometry and unseen space, and rasterizes the resulting structured noise into arbitrary views. Because rendering can introduce local correlations, we further propose a cross-view-consistent whitening procedure that decorrelates neighboring pixels while preserving correspondence across views. The structured noise serves as the diffusion prior for a LoRA-adapted pretrained video diffusion model, requiring no architectural modifications. By enforcing spatiotemporal and cross-view coherence at the source of stochasticity, SyncVid generates synchronized videos and improves both perceptual quality and cross-view consistency in our experiments. Project page: https://qualcomm-ai-research.github.io/SyncVid-Webpage.

Figures & tables

Figure 1
  View SynchronizationTemporal ConsistencyCamera Accuracy
MethodGeneration LengthMat. Pix. ↑\uparrowFVD-V ↓\downarrowCLIP-V ↑\uparrowCLIP-F ↑\uparrowRotErr ↓\downarrowTransErr ↓\downarrow
  DynamicStatic     
DaS [ 22 ]494941251315.6986.0197.400.15490.2164
Go-with-the-Flow [ 11 ]496871277296.3087.4496.780.12250.1109
ReCamMaster [ 4 ]7510731428233.7088.0697.390.16650.1778
ArtiFixer-14B [ 14 ]7511281937195.5588.9997.550.07800.0529
Table 1: Quantitative comparison on view synchronization, temporal consistency and camera accuracy on Multi-ViewBench videos. Metrics are defined in Sec. 4.2 .
Table 2: Quantitative comparison on VBench [ 21 ] metrics .
Figure 2: Qualitative comparison on a Multi-ViewBench example with an Arc-right trajectory.
VariantView SynchronizationTemporal Consistency
  Mat. Pix. ↑\uparrowFVD-V ↓\downarrowCLIP-V ↑\uparrowCLIP-F ↑\uparrow
Noisestandard222518.7788.7697.34
w/o whitening208468.3988.5396.92
 w/ whitening267461.4491.0597.61
Conditioning+none (I2V)224830.0087.5397.80
+render view231574.5089.3497.59
Table 4: Ablations. Visual-quality and view-synchronization metrics on various factors.
Appendix figures & tables9 assets

Supplementary material from the paper’s appendix.

Appendix

FieldRangeRequirement or meaning
Gates (all must be true)
is_single_continuous_shotboolNo cuts or transitions.
is_real_time_playbackboolNo time-lapse, hyper-lapse, or slow motion.
is_physically_plausibleboolNo VFX blends, text overlays, or screen-in-screen.
is_motion_blur_acceptableboolNo motion blur severe enough to destroy structure.
is_focal_blur_acceptableboolNot excessively out of focus.
Appendix
Table 5: Stage-1 judgement schema. The model must return every field. Gates are hard requirements: a clip is admissible only if all six are true. The composition bucket defines the standard / complex split, and the semantic scores feed the richness index of Sec. A.1.3 .
#StageSurvivingRemoved
 Stage-1 judgements available7,561 
1Gate pass (all six gates true)6,617944 gate failures
2Landscape (width >> height)5,973644 portrait or square
3De-duplication (unique clips)4,5411,432 copies of 1,197 clips
4Moving camera body2,2312,310 static-source
5Optical-effect removal2,2274 optical-effect
Appendix
Table 6: Curation funnel. Counts at the time the benchmark was frozen, when 7,561 of the 9,194 corpus clips had been scored. De-duplication runs after the gate and landscape filters but before static-source classification; see Sec. A.1.5 for why the order matters.
 StandardComplex
Composition bucketsimplecomplex
Candidates in pool4041,823
Selectiontop 100 by ρ\rhobalance to 1,000, then top 100 by ρ\rho
ρ\rho range24.0–64.064.0–100.0
ρ\rho mean30.971.2
ρ\rho median32.064.0
Appendix
Table 7: The two released splits of Multi-ViewBench. Both hold 100 clips, are disjoint, and are drawn from the 2,227-clip pool of Table 6 . Richness ρ\rho is defined in Eq. 20 . Attribute rows give the distribution over the 100 clips.
Figure 3: Schematic of the six synthesized evaluation trajectories ; Every trajectory is a camera-local SE(3) delta Δt\Delta_{t} composed after the source camera’s pose at each frame (Eq. 21 ), so all six depart from the same source pose ( Start ) at t=0t=0 and diverge from it over the clip. Arc left/right (violet) sweep the camera around a pivot ahead of it in the local horizontal (right–forward) plane; Dolly in/out (orange) dolly along the local forward axis toward/away from that pivot; Pedestal-up-Tilt-down and Pedestal-down-Tilt-up (blue) shift the camera along its local down axis while tilting about its local right axis so the same pivot stays centred in frame.
 4D representation (NoPo4D)Noise rasterizationWhiteningVideo generation (WAN2.2-14B)
    DiTVAE decoderTotal
Compute time (sec.)4.186.440.5323.97 / step4.601203.1
Appendix
Table 8: Compute time breakdown for the four main components of SyncVid. Times are reported in seconds per input clip with 75 frames at 480 ×\times 832 resolution on an Nvidia H100 GPU. The DiT time is reported per denoising step.
Figure 4: Qualitative comparisons under a Dolly-out camera trajectory. For each example, six equally spaced frames are shown for the source video and each method’s generated reshoot.
Figure 5: Qualitative comparisons under an Arc-right camera trajectory. For each example, six equally spaced frames are shown for the source video and each method’s generated reshoot.
Figure 6: Qualitative comparisons under an Arc-left camera trajectory. For each example, six equally spaced frames are shown for the source video and each method’s generated reshoot.
Figure 7: Failure examples. Each example shows six equally spaced source frames in the first row, followed by the corresponding generated frames. The top example follows Arc-left and the bottom example shows Dolly-out trajectory.
More in this topic

Multi-View Video Generation

Multi-view video generation creates synchronized, viewpoint-consistent videos for controllable scene synthesis, reconstruction, and simulation.

See all 57 papers

Explore similar work

CardsList
Jan 23, 2026

AnyView: Synthesizing Any Novel View in Dynamic Scenes

Modern generative video models excel at producing convincing, high-quality outputs, but struggle to maintain multi-view and spatiotemporal consistency in highly dynamic real-world environments. In this work, we introduce AnyView\textbf{AnyView}, a diffusion-based video generation framework for dynamic view synthesis\textit{dynamic view synthesis} with minimal inductive biases or geometric assumptions. We leverage multiple data sources with various levels of supervision, including monocular (2D), multi-view static (3D) and multi-view dynamic (4D) datasets, to train a generalist spatiotemporal implicit representation capable of producing zero-shot novel videos from arbitrary camera locations and trajectories. We evaluate AnyView on standard benchmarks, showing competitive results with the current state of the art, and propose AnyViewBench\textbf{AnyViewBench}, a challenging new benchmark tailored towards extreme\textit{extreme} dynamic view synthesis in diverse real-world scenarios. In this more dramatic setting, we find that most baselines drastically degrade in performance, as they require significant overlap between viewpoints, while AnyView maintains the ability to produce realistic, plausible, and spatiotemporally consistent videos when prompted from any\textit{any} viewpoint. Results, data, code, and models can be viewed at: https://tri-ml.github.io/AnyView/

Basile Van Hoorick, Dian Chen, Shun Iwase +7

Toyota Research Institute, CA, USA · Amazon Web Services, CA, USA

Multi-View ConsistencyCamera-Conditioned Video GenerationVideo Diffusion ModelsMulti-View Video GenerationDynamic Novel View SynthesisTemporal Consistency in Video Generation+1
AskPDF
Jul 6, 2026

MV-Forcing: Long Multi-View Video Generation via 4D-Grounded Spatio-Temporal Self-Forcing

Recent advances in video diffusion models have enabled either long single-view generation through temporal autoregression, or short multi-view synthesis through bidirectional attention. However, generating long, multi-view consistent videos of dynamic scenes remains unsolved. In this work, we present MV-Forcing, a framework that composes temporal and view-wise autoregression within a single diffusion model by introducing a 4D geometric bridge between sequentially generated views. Our key insight is that an autoregressive 3D reconstruction model naturally interfaces between autoregressively generated views. Given a completed source view, we reconstruct its 3D structure and render a geometric prior of the next target viewpoint, which the diffusion model refines into a high-quality video. To extend generation beyond the teacher's fixed temporal window, we introduce a joint denoising regime where both view slots are initialized from noise during training, enabling temporally unbounded generation. We distill the model via Distribution Matching Distillation with Spatio-Temporal Self-Forcing, closing the train-inference exposure bias gap for both temporal and view-sequential autoregression. Extensive experiments on both synthetic and real-world data demonstrate that MV-Forcing produces geometrically consistent multi-view videos of dynamic scenes at arbitrary lengths and viewpoint counts using a single few-step student model.

Gal Fiebelman, Hadar Averbuch-Elor, Sagie Benaim

The Hebrew University of Jerusalem · Cornell University

Multi-View ConsistencyVideo Diffusion ModelsMulti-View Video GenerationLong Video GenerationAutoregressive Video GenerationMulti-View Diffusion Models
AskPDF
Jul 2, 2026

Track the Noise, Move the World:3D-Grounded Motion-Consistent Noise for Controllable Video Generation

Modern image-and-text-to-video diffusion models can synthesize highly realistic videos by iteratively denoising an initial Gaussian noise tensor conditioned on reference image and text inputs. However, existing approaches still lack precise and unified controllability over both object motion and camera motion within a single generation process. We present UniCaMo, a unified framework that enables simultaneous control of object trajectories and camera viewpoints by directly constructing the input noise of the diffusion model. Specifically, UniCaMo builds a shared 3D-grounded motion-consistent noise space across latent video frames. Sparse 3D point tracks are used to warp the Gaussian noise of the reference frame along desired object trajectories, while a virtual spherical noise representation provides globally consistent noise values for newly revealed scene regions under camera motion. By combining local track-guided noise warping with global sphere-based noise sampling, UniCaMo maintains geometric and temporal consistency under both object movement and viewpoint changes. Because UniCaMo modifies only the input noise, it requires no auxiliary adapters, control branches, or architectural changes to the underlying video diffusion model. With lightweight LoRA fine-tuning on large pretrained video diffusion models, including Wan 2.1 (14B), UniCaMo achieves state-of-the-art results in both video quality and motion controllability on standard controllable video generation benchmarks.

Long Vu, Tan Ngo, Animesh Karnewar +5

Qualcomm AI Research · Trinity College Dublin, Ireland

Camera-Controlled Video GenerationVideo Diffusion ModelsDiffusion Models for Motion GenerationControllable Video GenerationTrajectory-Guided Video Generation3D-Aware Video Generation
AskPDF