Gemini Omni has no 1:1: cut the square Google Ads wants with Sume
Asset Studio exports 16:9, 9:16 and square, but Gemini Omni on Sume renders only 16:9 or 9:16. Crop the 1:1 with video-filter: x 0, y 0.21875, 1 by 0.5625.

If you generate with Gemini Omni through Sume and need a square for Google Ads, render the clip at 9:16 or 16:9 and crop it to 1:1 with the video-filter model. Sume's gemini-omni-flash-1.1 accepts only 16:9 or 9:16, so there is no square option to request. For a 1080x1920 clip, a crop of x 0, y 0.21875, width 1, height 0.5625 returns a centered 1080x1080 frame.
Google's own announcement describes the opposite situation: Gemini Omni in Google Ads (read 2026-10-04) lists 16:9, 9:16 and square outputs inside Asset Studio, rolling out globally as a free feature. If you want the same three shapes from an API you control, you build the square yourself.
What does each side actually offer?
The table compares the vendor page with the Sume model page. Sume's list covers only gemini-omni-flash-1.1; other video models on the router list their own ratio subsets in supported_aspect_ratios.
| Where | Ratios | Length or note |
|---|---|---|
| Google Ads Asset Studio | 16:9, 9:16, square | Free feature, rolling out globally, per Google's page |
| Sume gemini-omni-flash-1.1 | 16:9 or 9:16 | 3-10 seconds, 360p to 4K, native audio always on |
| Sume video-filter crop | Any fraction you set | Width and height from 0.05 to 1, $0.02 per encode |
How do I get a 1:1 square from a 9:16 Omni clip?
Import the finished clip, then call video-filter with one crop op. The crop fields are fractions of the source frame: x and y in 0 to 1, width and height in 0.05 to 1, with x plus width and y plus height both at most 1. The docs do not promise automatic centering, so you set the offsets yourself.
For a 9:16 source the square keeps the full width and a 0.5625 slice of the height (1080 divided by 1920). Centered, the top offset is (1 minus 0.5625) divided by 2, which is 0.21875. For a 16:9 source, flip it: height 1, width 0.5625, x 0.21875, y 0.
{
"video_url": "https://media.sume.com/<your-artifact>",
"ops": [
{ "op": "crop", "x": 0, "y": 0.21875, "width": 1, "height": 0.5625 }
],
"mode": "sync"
}What should I check before uploading the square?
A center crop removes the top and bottom of a vertical frame, so write the prompt with the subject in the middle band. Keep any burned-in text out of the outer 22 percent at the top and bottom, or the crop will clip it.
The source must be a media.sume.com artifact, so a clip made by Sume already qualifies; an outside file needs a media import first. The call needs an Idempotency-Key header, and output inherits the source frame rate and audio unless the program changes them. Because Omni's audio is always on, the square keeps its soundtrack.
When is the free Asset Studio enough?
If you work only inside Google Ads, Asset Studio's multi-format export covers it and costs nothing per clip. The API route makes sense when the same clip also has to ship to channels Google Ads does not reach, or when you want every variant produced by one scripted run. The related hook variant bulk run post covers that side.
Provenance is a separate question. Google says assets from its tools carry imperceptible SynthID watermarking, and the stored post on SynthID and C2PA in Google Ads covers what that means for files you upload yourself.
Sources
Related posts
More in Models
- Genjutsu 30 reference images vs Sume's 1 to 8
Higgsfield's Genjutsu takes up to 30 reference images. Sume's higgsfield-genjutsu takes one source video plus 1 to 8 reference images, at 480p or 720p.
- Google Lyria ids and prices next to the Sume Music Router ids
Google lists Lyria 3.5 at $0.08 a song, a 30-second clip model at $0.04, and Lyria 3 Pro Preview at $0.08. Which of them Sume routes, and at what flat price.
- Google media model ids on Oct 4: Omni, Veo 3.1, Lyria 3.5 vs Sume ids
Which Google model id to call on Oct 4, 2026 and the matching Sume id: Omni Flash 1.1, Veo 3.1 previews, Lyria 3.5 and Lyria previews, with status and dates.
- GPT Image aspect ratios: Firefly's 10 vs Sume's 17
Firefly's September 2026 update gives GPT Image 2 ten aspect ratios instead of three. Sume's image API documents 17 ratios plus auto, and custom pixels.
Written by Sume