Reframe 16:9 to 9:16 before you burn captions, not after
Crop a 16:9 clip to 9:16 with video-filter (width 0.3164, x 0.3418), then caption. Order trim, crop, caption: $0.02 + $0.02 + $0.20 = $0.24.

Reframe before captions. A 9:16 window of a 16:9 frame is 81/256 = 0.3164 of the width at full height, centered at x = 0.3418, and captions burned onto the 16:9 file would be cropped away or sit off the safe area, so crop first and caption the vertical file. Trim, crop and caption cost $0.02 + $0.02 + $0.20 = $0.24 for a clip up to 60 seconds.
YouTube says Shorts are vertical videos up to 3 minutes with a 1080p maximum upload resolution (read 2026-10-09), so a 1080-high crop of a 1080p source is the right size.
The crop fractions
Video filter crops by fractions of the frame, so you do not need pixels. For source width W and height H, the 9:16 window has height 1 and width = (9/16) x H / W. For 1920 x 1080 that is 0.3164 (607.5 px of 1920), and x = (1 - width) / 2 = 0.3418. The compiler rounds to even pixel values for yuv420p.
{
"video_url": "https://media.sume.com/artifacts/artf_demo/talk-169.mp4",
"ops": [{ "op": "crop", "x": 0.3418, "y": 0, "width": 0.3164, "height": 1 }]
}Order and cost
| Step | Why this order | Rate |
|---|---|---|
| Video trim | Shortest file for every later job | $0.02 |
| Video filter crop | Vertical frame before text is placed | $0.02 |
| Video captions | Burned on the final geometry (up to 60 s) | $0.20 |
| Total | $0.24 |
Gotchas
- A center crop is a guess about where the speaker stands; extract one still with video frames before you commit.
/v1/video-filter/checkis free and reports diagnostics instead of a 400, so run it first.- Out-of-range fractions fail with
video_filter_crop_out_of_bounds; a side under 0.05 also fails. - The filter source must be 300 s or shorter.
Sources
Related posts
More in Media tools
- RMBG on a 512 px icon or a 4K photo: $0.0225 either way
Sume RMBG 1.0 charges $0.0225 per image and the catalog says the price does not vary by image size. A 1,000-image batch is $22.50 whatever the pixel counts.
- Remove backgrounds from 2,000 marketplace images for $45 with RMBG
Sume RMBG 1.0 is $0.0225 per image regardless of size, so 2,000 cutouts cost 2,000 x $0.0225 = $45.00. Request body, endpoint and limits.
- Render a 10-minute video with Sume Timeline: $1.00
Sume Timeline 1.0 charges $0.10 per output minute, rounded up: a 10-minute audio spine is $1.00. Maximum 1,800 seconds per render. Read 2026-10-09.
- Render a 15-minute video with Sume Timeline: $1.50
Sume Timeline 1.0 charges $0.10 per output minute, rounded up: a 15-minute audio spine is $1.50. Maximum 1,800 seconds per render. Read 2026-10-09.
Written by Sume