Depth video converter
Turn a short clip into a grayscale depth map video. Convert and download for free, or bring the result into a character template as a motion reference.
Original video and depth video, side by side
Compare the same coffee scene before and after depth conversion: the original RGB video and its estimated depth video. Play both clips to inspect how the scene's shapes and movement appear without its colors and textures.
1. Original RGB video
The 15-second color clip supplies the action: a mug moves while the other person keeps adding ingredients. Its actors, clothing, colors and background are still visible. This is the same input loaded by Try the coffee clip.
2. Grayscale depth video
Depth estimation turns the scene into changing near and far regions. Watch the people, mug and counter in shades of gray. Original color and texture are removed, leaving a simpler spatial reference to use alongside character images.
The depth video preserves a changing estimate of near and far surfaces. Compare its silhouettes with the color clip: small props, fingers and facial details can be simplified or lost. Inspect those areas before using the depth video as a motion reference in another tool.
The exported depth video is silent. Keep your original audio separately if you need it for an edit. Character-video generation and soundtrack options belong to the coffee template; they are separate from this free conversion.
Original action: Mr and Mrs Stretchy on TikTokWhat is a depth video?
Brightness represents estimated relative distance: nearby surfaces are lighter and distant surfaces are darker. The tool estimates depth across the clip and exports a grayscale MP4 with the same duration. This is depth estimation, not a black-and-white color filter.
Relative depth describes which surfaces appear nearer or farther within an image. The brightness is not a measurement in meters. A light area may be a nearby hand, face or object, even if that surface was dark in the color video. Depth maps can make the scene’s spatial layout easier to inspect without its original textures.
For a character remake, the depth video provides a view of the scene’s shape and movement, while your character images provide the intended appearance. Removing RGB detail can help you prepare a different reference, but the finished look still depends on the video generator and its interpretation of all the inputs.
How to convert video to a depth map
Choose your short video
Upload an MP4, MOV or WebM clip up to 15 seconds and 50 MB, or select the coffee example above. Preview the source to check that it contains the full action you want. Trim a longer video before choosing it here.
Convert and inspect the depth
Click Convert to depth. The first conversion downloads the model; keep this tab open while it runs. Watch the progress, or cancel if you need to stop. Compare the resulting depth video with the original, especially around hands, small props and overlapping people.
Download or continue creating
Save the silent depth MP4 to your device. To make a character version, click Use in coffee template after conversion. That transfers the converted video into the template; you choose the character images and review the quote before starting a separate generation.
Use your own characters
Send the result to the coffee template, add two character images, and review the generation quote. Depth guides shape and motion; the images provide character appearance. Check the generated result for the style and movement you want.
In the coffee example, check the missed scoop, the moved mug, the water and the milk before judging the remake. A recognizable cast alone does not tell you whether the joke still works. You can also keep the downloaded depth file for another tool that accepts video references.
Ways to use your depth video
Prepare a character motion reference
Start with a short scene whose action you can see clearly. Convert it to depth, then use separate character images for the cast. The coffee template provides a concrete two-character workflow; your own clip still needs to suit the scene and instructions you use there.
Inspect movement without color detail
Compare the original and depth video to see how figures overlap and objects move through the scene. This can help you spot areas where the estimated depth loses detail before spending credits on a character generation. It is a preview of relative structure, not a reconstructed 3D scene.