Dim a bright restaurant clip for captions: $0.02 plus $0.20

To make captions readable over a bright restaurant clip, dim the video first with Sume's Video filter, then burn the captions on the dimmed file. The filter costs $0.02 and the captions $0.20, so $0.22 for the pair (read 2026-10-06). The filter has a free check endpoint, so you can validate the request before paying.

Order matters. Captions are burned into the pixels, so a dim applied after captioning would darken the text too.

The two steps

Video filter takes a media.sume.com URL and a list of operations, up to eight per request, including dim and crop. Send one dim operation, run the check endpoint to confirm the request is valid, then submit it with an Idempotency-Key.

Send the resulting file to Captions. A clip with no speech needs the cues field, where you write the text and timing yourself.

Why not just use a text box

Bright food footage, such as a lit counter or a white plate, makes white text hard to read. A mild dim raises contrast without a design change. It is also repeatable: the same request works on every clip in a menu series, so each one looks consistent.

A heavy dim makes the food look dull, which defeats the clip. Try a light setting on one clip, look at it on a phone in daylight, and then apply it to the rest.

A check before you scale

Run the whole pair on one clip and compare it with the undimmed version on a phone. Look at the darkest and the brightest frames, because captions fail in the bright ones.

When you are happy, apply the same request to the other clips in the series. At $0.22 each, ten clips cost $2.20, and every one carries the same look. Keep the request body in your own notes so a later menu change can reuse it exactly.

Tradeoffs

The dim affects the whole frame, not only the area behind the text. If you want the food at full brightness, a caption design with its own background may be a better fit, and the captions design options can change style and placement.

Two jobs mean two files and two receipts. Keep the dimmed intermediate if you plan to caption the same clip in several versions, because you pay for the filter only once.