One 45-second ad, three sizes: caption once ($0.40), not thrice

The question

You have a finished 45-second 9:16 ad and need it at 9:16, 4:5 and 1:1 for Instagram Feed, Facebook Feed and a square placement. Captioning each size separately costs $0.90 at the Sume prices below. Captioning once at 9:16 and cropping two further renders from the captioned file costs $0.40, but only if the crop keeps the caption inside the frame.

The ledger

Prices come from the Sume docs: a Timeline render is $0.10 per output minute rounded up, so 45 seconds is 1 billed minute, and a standalone caption job is $0.20 for videos of up to 60 seconds (docs read 2026-10-09).

Will the captions survive the crop?

Timeline's video[].fit takes cover (default), contain, stretch or blur. Converting 9:16 to 4:5 with cover keeps the full width and shows 0.5625 / 0.8 = 70.3% of the height; to 1:1 it keeps 56.25% of the height. A caption centered near the middle of the frame survives both, because the crop removes the top and bottom edges equally. A caption anchored low does not.

The Meta Feed guidance is a reminder that the edges are crowded anyway: Instagram Feed shows 9:16 at 1080x1920 and Facebook Feed recommends 4:5 at 1440x1800 (pages read 2026-10-09). With contain the whole 9:16 frame is kept inside the new canvas with bars or blur, so no caption is cropped, at the cost of a smaller picture.

Settings for the two renders

Output width and height must be even integers from 256 to 2160, so 1440x1800 and 1080x1080 are both valid. Keep output.fps equal to the source frame rate to avoid the resampling warning the docs describe, or omit it so the render follows the source.

• 4:5 render: output: { width: 1440, height: 1800 }, fit: cover.
• 1:1 render: output: { width: 1080, height: 1080 }, fit: cover if the caption is central, otherwise contain.
• Check each result visually; the docs promise geometry, not that your words stay in frame.

When to pay for three caption jobs

Caption each size separately if the 1:1 or 4:5 version needs different wording, a different language or a different style. Otherwise the single caption is the cheaper and more consistent path.

Scaling the arithmetic

The saving grows with the number of sizes and ads. For 10 ads in three sizes, captioning every render is 10 x $0.90 = $9.00, while captioning each master once and rendering the two other sizes is 10 x $0.40 = $4.00. The difference is $5.00, which is small money against the cost of a caption that falls outside a crop, so spot-check at least the first ad of each batch on a phone-sized preview.

The comparison assumes each ad is 60 seconds or less, because the docs state the $0.20 estimate for clips of that length. A longer clip is outside what the docs price, so confirm the estimate before you queue it.