YHLUO DEMO / LYHDEMO · MUNDO PIPELINE CHECK

From raw frames to video.

Scenes of the Mundo project page, section 1.1, indoors and outdoors, each taken once through the deployment pipeline: the raw frames are reconstructed into a proxy scene, a camera path is designed from the scene, the proxy is rendered along it, and the video is generated. The number of chunks (one to three) and both seeds were fixed before the runs, and no result was picked from several.

Stage by stage

Each scene shows what the stages wrote, in order: the input frames, the reconstruction as a point cloud with the boxes of the exported scene (drag to orbit, scroll to zoom, pick an object to highlight its box), the proxy video next to the generated video on the same camera path, and for every chunk the context photographs chosen and the description written for the prompt.

Input frames

input frames

Reconstruction: point cloud and boxes

Full viewer with frame overlays →

Camera path: proxy and generated video

ProxyGenerated
The files as written: proxy.mp4 · output.mp4

Context photographs and written prompt per chunk

The same office space with a blue door, light wood desks, whiteboards, and shelving units against white walls, under bright overhead lighting, with a window bank and blinds on one side. The whole path: a smooth camera move in which the camera pulls back 0.6 m and sinks 0.4 m and turns right 15 degrees, then glides forward 1.2 m and slides left 0.9 m and turns right 15 degrees.

Chunk 1 · frames 0–123 · coverage 0.87
context photographcontext photographcontext photographcontext photographcontext photographcontext photograph

The camera pulls back slightly, sinks lower, and turns right, revealing more of the right workstation desktop and the left wall whiteboard, while the window bank with blinds remains visible on the left side of the frame. The scene is brightly lit with even overhead lighting, creating a clean, neutral atmosphere in the office space.

Chunk 2 · frames 90–213 · coverage 0.87
context photographcontext photographcontext photographcontext photographcontext photographcontext photograph

The camera glides forward, passing the left workstation's monitor and keyboard, then slides left to reveal the window bank with blinds and turns right to show the whiteboard on the right wall, with a high-back office chair visible in the background. The scene is brightly lit with even overhead lighting, creating a clean, professional atmosphere with no shadows or reflections.

Input frames

input frames

Reconstruction: point cloud and boxes

Full viewer with frame overlays →

Camera path: proxy and generated video

ProxyGenerated
The files as written: proxy.mp4 · output.mp4

Context photographs and written prompt per chunk

The same small bedroom with herringbone wooden floors, a bed with a blue headboard and gray duvet, white shelving units cluttered with items, a window with sheer curtains, and a backpack on the floor. light streams through the window. The whole path: a smooth camera move in which the camera slides right 0.7 m and glides forward 0.5 m and turns left 54 degrees, then slides right 0.8 m and sinks 0.6 m and turns left 35 degrees.

Chunk 1 · frames 0–123 · coverage 0.83
context photographcontext photographcontext photographcontext photographcontext photographcontext photograph

The camera slides right, glides forward, and turns left, passing the window with sheer white curtains, revealing the white shelving unit cluttered with items on the left, and then showing the black backpack resting on the wooden floor near the bed. Soft, natural light streams through the window, illuminating the room with a calm, quiet atmosphere.

Chunk 2 · frames 90–213 · coverage 0.69
context photographcontext photographcontext photographcontext photographcontext photographcontext photographcontext photographcontext photograph

The camera slides right, sinks lower, and turns left, passing close to the window curtains, revealing the white shelving units cluttered with items, and eventually showing the black backpack resting on the wooden floor near the bed. Soft, natural light streams through the sheer curtains, casting gentle illumination across the room and creating a calm, still atmosphere.

Input frames

input frames

Reconstruction: point cloud and boxes

Full viewer with frame overlays →

Camera path: proxy and generated video

ProxyGenerated
The files as written: proxy.mp4 · output.mp4

Context photographs and written prompt per chunk

The same grand, circular interior hall with marble columns, wooden benches, and a central mosaic floor, illuminated by warm artificial lights and natural light from clerestory windows, featuring a second-floor gallery with ornate railings. The whole path: a smooth camera move in which the camera slides right 1.5 m and turns left 15 degrees.

Chunk 1 · frames 0–123 · coverage 0.79
context photographcontext photographcontext photographcontext photographcontext photographcontext photograph

The camera slides right and turns left, passing the wooden bench on the left, revealing the main stair flight to the right, and moving past the central information stand and marble columns with gold capitals as it turns. Warm artificial lights and natural light from clerestory windows illuminate the grand hall, casting a soft glow on the marble columns, wooden benches, and central mosaic floor.

Input frames

input frames

Reconstruction: point cloud and boxes

Full viewer with frame overlays →

Camera path: proxy and generated video

ProxyGenerated
The files as written: proxy.mp4 · output.mp4

Context photographs and written prompt per chunk

The same ornate church interior with rows of wooden pews, a high altar with religious artwork, arched doorways with red curtains, and a painted ceiling with stars and angels, lit by warm artificial lights and natural light from tall windows. The whole path: a smooth camera move in which the camera glides forward 4.7 m and slides right 0.7 m and turns right 15 degrees, then glides forward 3.4 m and slides left 3.2 m and turns right 91 degrees, then pulls back 3.9 m and slides left 1.1 m and turns right 79 degrees.

Chunk 1 · frames 0–123 · coverage 0.66
context photographcontext photographcontext photographcontext photographcontext photograph

The camera glides forward 0.6 meters, slides left 0.2 meters, and turns left 10 degrees, revealing the high altar on the right side of the frame, which is covered in white cloth and has a wooden altar table, while the raised sanctuary platform and front steps are visible on the left side of the frame. The camera glides forward 4.1 meters, slides right 0.7 meters, and turns right 25 degrees, passing the presider chair and middle nave chandelier, revealing the left front seating row and right front seating row, and eventually focusing on the left front seating row which is close and large in the frame. The scene is lit by warm artificial lights and natural light from tall windows, creating a serene and reverent atmosphere.

Chunk 2 · frames 90–213 · coverage 0.58
context photographcontext photographcontext photographcontext photographcontext photograph

The camera glides forward past rows of wooden pews with light-colored cushions, then slides left and turns right to reveal the high altar with religious artwork and statues, framed by arched doorways with red curtains, all under a painted ceiling with stars and angels. The scene is illuminated by warm artificial lights and natural light from tall windows, creating a serene and reverent atmosphere.

Chunk 3 · frames 180–303 · coverage 0.62
context photographcontext photographcontext photographcontext photographcontext photograph

The camera slides left and turns right, passing the left wall lantern and revealing the left seating rows with their wooden pews and light-colored cushions, while the red curtains of the curtained bay door remain visible in the background. The camera pulls back and slides left, revealing more of the left seating rows and the front left chapel devotional image, with the warm artificial lights illuminating the wooden pews and the red curtains of the bay door. The camera pulls back further and slides right, revealing the raised sanctuary platform, the central architectural reredos with religious artwork, and the painted ceiling with stars and angels, while the right seating rows with their wooden pews and light-colored cushions come into view. The scene is lit by warm artificial lights and natural light from tall windows, creating a serene and reverent atmosphere in the ornate church interior.

Input frames

input frames

Reconstruction: point cloud and boxes

Full viewer with frame overlays →

Camera path: proxy and generated video

ProxyGenerated
The files as written: proxy.mp4 · output.mp4

Context photographs and written prompt per chunk

The same historic abbey complex perched on a rocky island, surrounded by a tidal flat and bordered by a fortified wall. its main masonry body and upper church feature stone construction with a prominent spire. the village below has gabled roofs and is nestled among green trees under clear blue skies, illuminated by warm golden light. The whole path: a smooth camera move in which the camera slides right 1.4 m and glides forward 1.2 m and turns left 9 degrees, then slides right 1.3 m and sinks 1.3 m and turns left 14 degrees, then slides right 1.6 m and pulls back 0.7 m and turns left 15 degrees.

Chunk 1 · frames 0–123 · coverage 0.73
context photographcontext photographcontext photographcontext photographcontext photographcontext photograph

The camera glides forward and slides right, passing the long front rampart and revealing the abbey's main masonry body and its dark spire, while also showing the front slope woodland to the right and the village's gabled roofs and trees below. The scene is bathed in warm golden light under a clear blue sky, with the abbey's light brown stone and the village's dark roofs and green trees clearly visible.

Chunk 2 · frames 90–213 · coverage 0.61
context photographcontext photographcontext photographcontext photographcontext photographcontext photograph

The camera slides right and sinks lower, revealing more of the front slope woodland on the right and the long front rampart on the left, while the Abbey's main masonry body remains central, its light brown stone and dark spire prominent, and the side slope woodland stays visible on the right. The scene is bathed in warm golden light under a clear blue sky, with the village's stone buildings and green trees nestled below the abbey.

Chunk 3 · frames 180–303 · coverage 0.41
context photographcontext photographcontext photographcontext photographcontext photographcontext photograph

The camera slides right, pulls back, and turns left, revealing the abbey's main masonry body and upper church with their light brown stone and dark spire, while passing the front slope woodland and long front rampart, and showing the side slope woodland and village terrace tree clump. The scene is illuminated by warm golden light under clear blue skies, with the abbey and village nestled among green trees.

Input frames

input frames

Reconstruction: point cloud and boxes

Full viewer with frame overlays →

Camera path: proxy and generated video

ProxyGenerated
The files as written: proxy.mp4 · output.mp4

Context photographs and written prompt per chunk

The same ornate rotunda with pinkish columns and a domed roof, surrounded by water and trees under a clear blue sky, with detailed carvings and statues visible on its structure. The whole path: a smooth camera move in which the camera slides right 5.3 m and pulls back 1.4 m and turns left 45 degrees.

Chunk 1 · frames 0–123 · coverage 0.56
context photographcontext photographcontext photographcontext photographcontext photographcontext photographcontext photographcontext photographcontext photograph

The camera slides right and rises, passing the northeast rotunda support and revealing the relief-covered projecting pedestal, all with pinkish columns and detailed carvings under a clear blue sky. The camera pulls back and turns left, revealing the southeast and south rotunda supports, the rotunda entablature and roof, and the surrounding trees and water. The camera slides right and sinks, passing the northwest rotunda support and revealing the west bank tree, with the rotunda's pinkish columns and detailed carvings visible. The camera slides right and sinks again, revealing the south bank tree and rear approach tree, with the rotunda entablature and roof centered in the frame. Bright sunlight illuminates the scene, casting sharp shadows and highlighting the warm tones of the stone and the vibrant blue of the sky and water.

Input frames

input frames

Reconstruction: point cloud and boxes

Full viewer with frame overlays →

Camera path: proxy and generated video

ProxyGenerated
The files as written: proxy.mp4 · output.mp4

Context photographs and written prompt per chunk

The same brick lighthouse station with a tower, adjacent museum hall, and flagpole under clear blue sky, featuring wooden benches, green-shuttered windows, and an american flag waving, with distant trees and water visible. The whole path: a smooth camera move in which the camera pulls back 2.9 m and sinks 0.2 m, then pulls back 2.2 m and slides right 2.1 m.

Chunk 1 · frames 0–123 · coverage 0.57
context photographcontext photographcontext photographcontext photographcontext photograph

The camera pulls back and sinks slightly, revealing more of the brick museum hall on the left, the lighthouse tower on the right, and the American flag on its pole in the center, with the flag waving in the breeze. The scene is bathed in bright, clear daylight under a blue sky, with the red brick buildings and green-shuttered windows sharply defined.

Chunk 2 · frames 90–213 · coverage 0.31
context photographcontext photographcontext photographcontext photographcontext photograph

The camera pulls back and slides right, revealing more of the brick museum hall on the left and the lighthouse tower on the right, with the American flag on its pole visible in the center of the frame. The camera passes the wooden bench and the flagpole, showing the red brick walls, green-shuttered windows, and white door of the museum hall, and the red brick lighthouse tower with its white lantern room and cap. The scene is bathed in bright, clear daylight under a vivid blue sky, with distant trees and water visible on the horizon.

Runs

One cold-start run per row, all times in seconds: the reconstruction on one GPU, the generation on one eight-GPU node. Wall clock is the sum of the two commands. Model loading is the captioner and the condition encoder; prepare is the context selection, proxy rendering, prompt writing and condition encoding of all chunks; sampling is the diffusion steps and decoding of all chunks. What remains of the wall clock is loading the video model and the hand-over between stages.

scenechunks / frameswall clockreconstructionpath designmodel loadingpreparesampling
Office2 / 21488444847141253
Bedroom2 / 21495650437843260
Rotunda museum1 / 124122392557533124
Mission church3 / 30416871068226975378
Mont-Saint-Michel3 / 3041147511576258390
Palace of Fine Arts1 / 124973613537632133
Brick lighthouse2 / 214982509427742250

Blocking issues are the objects the reconstruction's final review still rejects. Clearance is the smallest distance between the camera path and a proxy surface. Coverage is the share of each chunk's view that its context photographs show, one value per chunk. Seeds are the path seed and the sampling seed.

sceneenvironmentobjectsblockingpath mclearance mcoverageseeds
Officeindoor2232.90.320.87, 0.873153 / 85727
Bedroomindoor2312.00.390.83, 0.690 / 42
Rotunda museumindoor11171.80.670.796034 / 46906
Mission churchindoor1801512.30.320.66, 0.58, 0.623322 / 91758
Mont-Saint-Micheloutdoor2956.01.430.73, 0.61, 0.419564 / 23265
Palace of Fine Artsmixed41116.00.320.562327 / 52015
Brick lighthouseoutdoor1455.91.350.57, 0.315114 / 31720

Settings