Skip to main content

Create a dolly zoom video

Updated September 8, 2026Try Dolly Zoom app

Dolly Zoom is a preset in Camera Movements, part of Workroom's AI workspace for video, image, and voice generation. It takes a single photo and generates a 5-second video with the effect Hitchcock's crew built for Vertigo in 1958. The camera pulls back while the lens zooms in, so your subject stays the same size in frame while the space behind them stretches and warps. One photo and one line of text, no footage, dolly track, or editing skills required.

Apps share a monthly pool of 5 free generations on the Free plan — once it's used, each run costs credits, with the exact cost shown on the Generate button.

Open Dolly Zoom

Open Apps from the sidebar and click Camera Movements. The Dolly Zoom preset is selected in the panel on the right, above the two input steps.

Camera Movements app with the Dolly Zoom preset selected and the Your photo upload panel
Camera Movements app with the Dolly Zoom preset selected and the Your photo upload panel

Add your photo

Under Step 1: Your photo, upload a JPG, PNG, or WEBP from your device. Or click Select to pick an image already in your Workroom library — anything you generated or uploaded earlier works.

Depth is what makes this effect read. The background is the part that stretches. A photo with real distance behind the subject — a corridor, a street, a room with a far wall — gives the move something to work with. A flat backdrop leaves less to bend.

Avoid backgrounds built from a repeating pattern that recedes to the center of the frame: a row of identical arches, a colonnade, a tunnel of matching doorways. The generator tends to read those as a passage and animate the walls themselves instead of changing the perspective.

Describe the shot

Under Step 2: What is in the shot, write one line in English describing what's in the frame — the subject first, then where it is and what it looks like: "a young woman with dark hair stands in a bright empty room."

Describe the scene, not the motion. Words like zooming, pulling back, or stretching fight the preset. The camera move is already built in, so the generator may add that motion on top and the effect comes out muddled. Stick to verbs of posture (sits, stands, leans) and words of place (behind her, along both walls). Keep the subject still for the same reason: the movement belongs to the camera.

Step 1 with an uploaded photo and Step 2 filled in with a one-line shot description
Step 1 with an uploaded photo and Step 2 filled in with a one-line shot description

Generate the video

Click Generate. There's no format setting to choose — the video comes out in the same aspect ratio as the photo you uploaded, so vertical stays vertical and widescreen stays wide.

The result is a 5-second, silent video you can play, download, or reuse in other Workroom tools.

Generated dolly zoom result — the subject holding its size while the corridor behind stretches, with the source photo and shot description in the panel
Generated dolly zoom result — the subject holding its size while the corridor behind stretches, with the source photo and shot description in the panel

Each generation comes out a little different, even with the same photo and shot description. If the warp feels too strong or too subtle, just generate again.

Troubleshooting

The background slides around instead of stretching

The source photo is probably built on a repeating pattern that recedes to the center — nested archways, a colonnade, a corridor lined with identical doors. Use a photo where depth comes from several different objects at different distances instead: a subject up close, a building further back, a horizon behind that.

The video moves in a way I didn't expect

Check your shot description for motion words. The camera pulls back, zooming in, the walls stretch — the generator treats these as instructions and adds that movement to the preset's own. Rewrite the line as a plain description of what's in the frame.

The effect is barely visible

There isn't enough depth in the photo. A subject shot against a wall, a plain sky, or a shallow backdrop gives the perspective almost nothing to change. Pick a photo with visible distance behind the subject.

Is this the same as Crash Zoom?

No. A crash zoom is a fast lens zoom with no camera movement — it snaps toward the subject and the perspective never changes. A dolly zoom moves the camera and the lens against each other, so the subject holds its size and the background warps. See Create a crash zoom video for that preset.

Can I use the result commercially?

Personal use is included on every plan; commercial use requires a Pro subscription. See Manage credits and billing for details on plans and credits.

Next steps

Dolly Zoom works well together with other Workroom apps: fix a low-resolution source with the Image Upscaler before generating, reframe your photo first with Change a camera angle, or score the finished video with Add sound to any video.