Estimate the depth a photograph never recorded, correct it by eye, and turn the result into moving images.
The estimate is a guess, not a measurement — it will sometimes place things at the wrong distance. These tools exist to make those errors visible and fixable.
In the field. Picks a depth pathway, walks each shot, and reads the scene for the clipping that breaks depth estimation.
A multiplane camera. Stand images as cards at chosen distances and photograph them — the depth map comes out exact, not estimated.
Open a photo, sweep the camera to expose depth errors, and correct them while it moves. Flags the tearing automatically.
Present the result with drift, orbit, flythrough, sway, or a configurable Hitchcock dolly zoom. Records video or frames.
Cut plates to a timeline and place the moments where the depth opens up. One intensity curve drives every camera move.
Synthetic scenes with known depth. Measures what the estimator gets right and wrong, instead of leaving it to the eye.
Everything the suite knows, searchable. A helper briefed on it, and proof plates scored for how strongly the effect will read.
Turn a frame sequence into a single video. MP4 where the browser can encode it, WebM everywhere else.
Everything runs in the browser. No image, depth map or video is ever uploaded — the files stay on your machine.
Each app carries a walkthrough on first visit and a ? button holding a full reference. Press ? at any point.
MP4 output needs WebCodecs, present in current Chrome, Edge, Firefox, Opera and Safari 26 or later. Where it is missing, the assembler produces WebM instead — the same picture in a different container.
Already have a depth map, or real geometry from a scan? Load it in the workbench or viewer directly and skip the estimate.