Loading video...

Video Failed to Load

Go Home

Guys, Depth Map Control now live on Mini Apps The workflow is quite easy: reference video -> depth map video -> reference image + depthvideo -> a video that follows motion and space more precisely. RT + Follow + Reply = 150 Credits in DMs (72H ONLY)

416,227 views • 3 months ago •via X (Twitter)

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

THE DEPTH MAP TRICK THAT FIXED DANCE ACCURACY IN SEEDANCE 2.0 Feed the model a video of someone dancing and it tries to interpret everything- the person, the clothes, the lighting, the room, and somewhere in there, the movement. Feed it a depth map and there's nothing left to interpret but the motion. Most creators trying to transfer a dance to a character reference the source footage directly, then wonder why the choreography drifts. The problem isn't the model - it's that you handed it ten variables when you only wanted one. Here's the workflow 1. Lock the character reference in GPT Image 2 first -face, build, costume, so identity holds independently of whatever motion gets applied to it 2. Convert the source dance footage into a depth map instead of using the raw video -this strips out the original performer's appearance, clothing, and environment entirely 3. Feed the depth map as the motion reference and the character sheet as the identity reference- two separate inputs doing two separate jobs, not one input trying to do both 5. Let the depth map carry only spatial movement -the model receives body position and momentum with no competing information about who's moving or what they look like 6. Keep the character and motion inputs isolated throughout - the moment you mix appearance data into the motion reference, the model starts negotiating between two identities Why this works • Raw footage passes the model everything at once- performer, wardrobe, room, lighting -and the choreography competes with all of it for attention • A depth map is pure spatial information, so the only thing left to transfer is movement • Separating identity from motion means the character can stay locked while the dance stays accurate - normally you're trading one for the other • The accuracy gain isn't the model getting better, it's the model getting fewer decisions to make Use cases: ⁃ Dance and choreography transfer onto original characters ⁃ Motion capture-style workflows without motion capture ⁃ Any sequence where a specific movement needs to survive intact ⁃ Character showcase content built on existing performance footage The character sheet answers who's dancing. The depth map answers how - and keeping those two questions separate is the whole trick.

Nexlow

86,365 views • 4 months ago

THE DEPTH MAP TRICK THAT FIXED DANCE ACCURACY IN SEEDANCE 2.0 Feed the model a video of someone dancing and it tries to interpret everything- the person, the clothes, the lighting, the room, and somewhere in there, the movement. Feed it a depth map and there's nothing left to interpret but the motion. Most creators trying to transfer a dance to a character reference the source footage directly, then wonder why the choreography drifts. The problem isn't the model - it's that you handed it ten variables when you only wanted one. Here's the workflow 1. Lock the character reference in GPT Image 2 first -face, build, costume, so identity holds independently of whatever motion gets applied to it 2. Convert the source dance footage into a depth map instead of using the raw video -this strips out the original performer's appearance, clothing, and environment entirely 3. Feed the depth map as the motion reference and the character sheet as the identity reference- two separate inputs doing two separate jobs, not one input trying to do both 5. Let the depth map carry only spatial movement -the model receives body position and momentum with no competing information about who's moving or what they look like 6. Keep the character and motion inputs isolated throughout - the moment you mix appearance data into the motion reference, the model starts negotiating between two identities Why this works • Raw footage passes the model everything at once- performer, wardrobe, room, lighting -and the choreography competes with all of it for attention • A depth map is pure spatial information, so the only thing left to transfer is movement • Separating identity from motion means the character can stay locked while the dance stays accurate - normally you're trading one for the other • The accuracy gain isn't the model getting better, it's the model getting fewer decisions to make Use cases: ⁃ Dance and choreography transfer onto original characters ⁃ Motion capture-style workflows without motion capture ⁃ Any sequence where a specific movement needs to survive intact ⁃ Character showcase content built on existing performance footage The character sheet answers who's dancing. The depth map answers how - and keeping those two questions separate is the whole trick.

Nexlow

120,346 views • 2 months ago

THE DEPTH MAP TRICK THAT ALSO WORKS ON EXPRESSION, NOT JUST DANCE — GPT IMAGE 2 × SEEDANCE 2.0 Feed the model raw footage of someone gesturing, reacting, moving through a scene, and it tries to interpret everything at once - the face, the outfit, the alley, the light, and somewhere in there, the actual motion. Feed it a depth map instead and there's nothing left to read but the movement itself. Most creators assume this trick only applies to dance, where the motion is clean and repetitive. It works just as well on loose, expressive, conversational movement - the kind with no fixed rhythm to lean on. The problem was never the model's ceiling. It's that raw footage hands it ten variables when the only one that matters is how the body is moving. Here's the workflow 1. Lock the character reference in GPT Image 2 first - face, build, outfit, so identity holds no matter how expressive or loose the motion gets 2. Convert the source footage into a depth map instead of feeding the raw video - this strips out the original performer's face, clothing, and environment entirely 3. Feed the depth map as the motion reference and the character sheet as the identity reference - two separate inputs, two separate jobs 4. Let the depth map carry only spatial movement - gesture, weight shift, head turn, with no competing information about who's moving or what they look like 5. Keep the two inputs isolated all the way through - the moment appearance data leaks into the motion reference, the model starts negotiating between two identities Why this works - Raw footage hands the model everything at once, and loose, expressive motion has nowhere near as much rhythm to lean on as dance does - A depth map is pure spatial information, so unscripted gesture reads just as cleanly as choreography - Separating identity from motion means the character stays locked while every twitch and turn still survives intact - The gain isn't a smarter model, it's a model with fewer decisions left to make Use cases: ⁃ Expression and gesture transfer, not just choreography ⁃ Conversational or reactive scenes with no fixed rhythm ⁃ Any footage where identity needs to change but the exact motion can't ⁃ Character showcase content built on found or reference footage The character sheet still answers who's moving. The depth map still answers how - it just turns out how applies to a lot more than dancing.

Primee32

30,594 views • 4 months ago