Subtitle position on video: Sume anchor_ratio for captions

To set where burned-in subtitles sit on a video, Sume's caption job takes design.placement with anchor_ratio and landscape_anchor_ratio. Each is the caption line's centre as a fraction of frame height. landscape_anchor_ratio is a second value for the landscape case, so one request can place text for both shapes.

Facts are from Sume's Video captions docs, read 2026-10-01. On the subtitle side of the news, Blackmagic's 2026-09-08 release says Cloud Presentations has been updated with subtitle support.

What do the two placement fields mean?

The docs define both as the line centre expressed as a fraction of frame height. The API schema accepts 0.05 to 0.95 for both and labels anchor_ratio as the portrait value. The docs do not say which edge the fraction counts from, so test a value; a number outside the range returns a 400 instead of a mis-placed render.

Where does each style put captions by default?

The docs say black-outline is a thick black outline "mid-frame", and korean-ad is lower-third. Overrides merge over the style, so sending only placement moves the text and leaves colours, weight and motion alone.

curl -X POST https://api.sume.com/v1/video-captions \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: caption-placement-001" \
  -d '{
    "video_url": "https://example.com/clean.mp4",
    "style": "black-outline",
    "design": { "placement": { "anchor_ratio": 0.8 } }
  }'

Does this work on every style?

No. design is not supported on punch or tiktok-green. It applies to slam and the Hangul identities. Details in the punch and tiktok-green note.

Can I re-place captions without transcribing again?

Yes. Pass source_caption_id instead of video_url with a new design. Sume reuses the source video and the word timings it has, so no second speech-to-text runs. The restyle is still a render and bills as one.