Square 1:1 video from Gemini Omni: it only makes 16:9 or 9:16

Omni outputs 16:9 or 9:16 only. For a 1:1 square feed post, render a Sume Timeline at 1080x1080 with fit cover or blur; the render is $0.10 per output minute.

4 min readSume
All posts

Gemini Omni Flash 1.1 outputs 16:9 or 9:16 only, so a square 1:1 post needs a second step. On Sume, generate the clip at 9:16 or 16:9, then render a Timeline with output.width and output.height both 1080 and fit set to cover (crop to fill) or blur (keep the whole frame). The render costs $0.10 per output minute on top of the Omni clip.

What each side supports

Google's Omni guide lists aspect ratios 16:9 and 9:16. Sume's Video Router accepts the same two for Omni. Timeline 1.0 accepts even output.width and output.height values from 256 to 2160, and each slot's fit is cover, contain, stretch or blur.

Getting 1:1 out of a 16:9 or 9:16 Omni clip (Google and Sume docs, read 2026-10-04)
Source clipTimeline fitResult in 1080x1080
16:9 landscapecoverCrops the sides; subject must be centered
16:9 landscapecontainWhole frame, with empty bars top and bottom
16:9 landscapeblurWhole frame; the blur mode fills the bars
9:16 verticalcoverCrops top and bottom; keep faces in the middle

Choose the source shape first

Cropping either shape to a square keeps 1080 of 1920 pixels along the long side, so about 44 percent of the frame is lost. Prompt for a centered subject and generate the shape that loses the least for your shot. A product on a table works at either.

Ask for the framing in the prompt ("centered, with space around the subject") because fit cannot recover what the model cropped in.

The cost and the steps

The Omni clip bills at provider list times 1.25 per second; a ten-second 720p clip is $1.25. The Timeline render adds one billable minute, $0.10. Import the clip through POST /v1/media-imports if it is not a Sume-hosted URL, call POST /v1/timeline-1.0/plan for an unbilled check, then render. Timeline needs an audio spine or audio.mode: "silence", so declare the length as the clip length.

Check the first render by eye. The blur look is a style choice; some brands prefer plain crop.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume