AI avatar video aspect ratio: 9:16, 1:1, 3:4, 4:3, 16:9

Sume's avatar video takes an aspect_ratio of 1:1, 3:4, 9:16, 4:3 or 16:9, and defaults to 9:16 when you omit it. Resolution is currently 720p, so the ratio is the only framing choice you make on size. Set it once at the top level of the request; on a preview it is locked in, and changing it means a new preview.

What shape is each ratio?

Where do I set it?

Put aspect_ratio next to avatar_handle in the POST /v1/avatar-1.0/talking-video body. The multi-scene example in the docs sets it at the top level, beside quality; there is no per-scene ratio in that example.

curl -X POST https://api.sume.com/v1/avatar-1.0/talking-video \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: avatar-ratio-001" \
  -d '{
    "avatar_handle": "my_presenter",
    "aspect_ratio": "16:9",
    "script": "Here is how the setup works in three steps."
  }'

Can I change the ratio after a preview?

Not without a new preview. The preview docs list aspect_ratio among the structural fields, along with script, video_inputs, avatar_handle and scene, that still require a new preview. Only quality can be overridden at the generate-video step, and only for the final render.

So decide the ratio before you review first-frame stills, not after. See avatar video first-frame previews.

What stays fixed whatever the ratio?

The window and tiers do not change with shape. Duration is still an estimated 4 to 60 seconds, quality is still standard, plus (the default) or max, and resolution is currently 720p for every ratio in the list.

If you need a ratio that is not one of the five, the docs list no other. Pick the nearest and crop in your own editor.

Does the ratio change the price or the length limit?

The avatar video page ties neither to the ratio: it lists the five values, the 720p resolution and the 4 to 60 second window as separate rules. Read the price for your exact request from the API instead of assuming one ratio costs less.

Inline captions are burned into the final MP4 after generation, so they follow whichever ratio you chose; they are not applied to preview stills.

Which ratio should I choose?

Match the place the video will play. The docs do not tie a ratio to a platform, so check the destination's own spec page instead of assuming one. Vertical feeds usually want portrait, a slide-style embed usually wants 16:9, and 1:1 is a compromise when a video must sit in a mixed layout.

If you use a photo scene reference, pick a photo close to the ratio you asked for; the docs do not describe how a mismatched photo is fitted, so check a preview.