MotionStream

Real-Time Video Generation with Interactive Motion Controls

Supplementary Material - Anonymous Submission

Best viewed in Chrome • Click videos for fullscreen

Real-time Interactive Streaming Demo on a single H100 GPU: Using simple drag and mouse motions, MotionStream enables real-time control of diverse scenarios including both object motion and camera movement with various grid configurations. Network latency plays an important role in streaming performance (some demo samples may exhibit slightly higher latency due to intercontinental connections). Our method maintains stable streaming with constant attention context. The system demonstrates robust long-video generation capabilities: by anchoring to the clean first chunk with a single attention window (keeping minimal drifted chunks within the attention context), the model often recovers quality even after disruptions, showcasing its resilience in long video streaming scenarios.
Cute Monster
Elephant
Border Collie
Shiba Dog
Anime
Tokyo Girl
Rocket Launch
Chess Demo
Long Streaming Demo
Notation: C = chunk size, S = number of sink chunks, W = local window size. Results demonstrate that maintaining at least one sink chunk is crucial for stable long-video generation. All experiments use our 1.3B causal distilled model, showing the effectiveness of our attention sink mechanism for extended temporal extrapolation.

Sample 1: Extrapolation Comparison

c3s0w1
c3s0w6
c3s1w1

Sample 2: Extrapolation Comparison

c3s0w1
c3s0w6
c3s1w1
Comparison with Other Baselines: We compare our methods with recent motion-controlled video generation approaches using samples from Sora demos. We denote our models as 1.3B-T (1.3B parameter teacher model), 1.3B-S (1.3B parameter student/distilled model), 5B-T (5B parameter teacher model), and 5B-S (5B parameter student/distilled model). Our methods consistently demonstrate high video quality and motion adherence compared to baseline approaches, with the student models achieving real-time performance.

Sample 1

Track Viz
GWTF
DAS
ATI
Ours 1.3B-T
Ours 1.3B-S
Ours 5B-T
Ours 5B-S

Sample 2

Track Viz
GWTF
DAS
ATI
Ours 1.3B-T
Ours 1.3B-S
Ours 5B-T
Ours 5B-S

Sample 3

Track Viz
GWTF
DAS
ATI
Ours 1.3B-T
Ours 1.3B-S
Ours 5B-T
Ours 5B-S
Motion Transfer: Given an initial image of similar structure, MotionStream can naturally transfer motions of arbitrarily long videos (can be done in real-time with online trackers).

Sample 1

Source
Motion Track
Result

Sample 2

Source
Motion Track
Result
Guidance Ablation Study: Using high motion guidance leads to overly rigid translations due to strict track adherence and ignores text cues (below sample is prompted with "rainbow appearing at the back"). While prompt guidance alone shows inferior quantitative metrics for motion reconstruction, it enables flexible text-based control. Our proposed joint guidance strategy effectively balances these two.

Sample 1

Motion Track
Motion Only
Joint Guidance
Prompt Only

Sample 2

Motion Track
Motion Only
Joint Guidance
Prompt Only