16:9 avatar video for slides and webinars: aspect ratio on Sume
Sume avatar video defaults to 9:16 but supports 16:9 and 4:3. How to set it for slides and webinars, with a 45-second Plus cost of $11.03 and caveats.
For a landscape slide or webinar intro, set aspect_ratio to 16:9 on the avatar video request: Sume supports 1:1, 3:4, 9:16, 4:3 and 16:9, defaults to 9:16, and currently renders at 720p only (read 2026-10-03). A 45-second 16:9 clip on Plus costs about $11.025 at the listed no-product rate.
The default is vertical because most avatar clips go to phone feeds. If you forget the field when making a lesson intro, you will get a tall video that needs letterboxing.
Choosing the ratio
The avatar video docs list the five ratios without stating a price difference, and the pricing page I read lists per-second rates by tier and product, not by aspect ratio. So choose by destination, not cost.
Use 16:9 for slide decks, webinar intros and web hero embeds; 4:3 for older slide templates and some LMS players; 9:16 for Shorts, Reels and TikTok; 1:1 and 3:4 for feed posts. For a talking head over slides, you may want the avatar in a corner, which is a Timeline composition task rather than an avatar setting.
| aspect_ratio | Typical use | Resolution |
|---|---|---|
| 16:9 | Webinar, slides, web hero | 720p |
| 4:3 | Legacy slide templates | 720p |
| 1:1 | Square feed post | 720p |
| 3:4 | Portrait feed post | 720p |
| 9:16 (default) | Shorts, Reels, TikTok | 720p |
A landscape request
Set the ratio explicitly and keep your script within 4 to 60 seconds. If the scene should look like a studio rather than a bedroom, direct it with a scene prompt, and keep the same prompt across a series so every lesson shares one set.
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: webinar-intro-001" \
-d '{
"avatar_handle": "webinar_host",
"aspect_ratio": "16:9",
"quality": "plus",
"scene": {"type": "prompt", "prompt": "Bright modern studio, soft key light, plain background"},
"script": "Welcome to today\u0027s session. In the next forty minutes we will walk through three ways to cut onboarding time in half."
}'What 720p means for slides
Because the resolution is currently fixed at 720p, a full-screen 1080p slide deck will have an avatar that is sharper at small sizes than at full-screen. Use the avatar as an inset, not a full-frame background, and keep text-heavy slides as separate images. If you need higher resolution for a final deliverable, check the docs for the current upscaling options rather than assuming.
Run a preview first. A landscape preview shows you whether the framing leaves space on the side for slide content, before you pay for the full render.
- Keep a consistent scene prompt for all lessons in a course.
- Use
captionsfor accessibility; the inline caption styles burn into the final MP4. - Test one slide and one clip together before producing the series.
Cost for a webinar series
A six-part webinar series with a 45-second host intro each is 270 seconds, which costs $49.68 on Standard, $66.15 on Plus and $148.50 on Max at the listed no-product rates (read 2026-10-03), plus $0.95 for the host avatar once. Because the aspect ratio does not appear in the rate card, you can switch between 16:9 for the webinar page and 9:16 for the social cut without changing the unit price, though each is its own render and its own bill.
That last point is worth planning for. A 16:9 clip cannot be re-cropped to a good 9:16 clip, since the avatar is framed for the wide shot. If you need both formats, budget two renders per intro, or use a preview for each to confirm the framing before paying for the full video.
Two checks before you render
Read the script aloud with a timer. Sume estimates duration from the text, and the estimate has to land between 4 and 60 seconds, so a 45-second intro should be roughly 110 words at a calm webinar pace. If you are near 60 seconds, split the intro into two jobs.
Then check the scene prompt for anything that clashes with a wide frame. A prompt that describes a desk close-up will look odd in 16:9. Describe the room, the light and the camera distance instead, and keep the description identical across the series.
Sources
Related posts
More in Models
- 16-second AI video API: only two Sume models take 16 seconds
One second past the 15-second ceiling, only Seedance 2.5 and Wan 3.0 accept 16 s on Sume. What a 16-second Wan clip costs at each resolution.
- 2:1 and 1:2 image models on Sume: banners and tall strips
Eight Sume image models list 2:1 and 1:2, from $0.025 to $0.075 per image. See which ones, what they accept for references, and a banner request.
- 20-second AI video API: which models take 20 s and what it costs
Only Seedance 2.5 and Wan 3.0 accept a 20-second request on Sume. Per-second rates for Wan, a worked cost, and the 400 the other rows return.
- 25-second AI video generator: which Sume model, and how much?
Seedance 2.5 and Wan 3.0 take 25 seconds in one request on Sume; every other prompt-driven row stops at 15. Wan per-second cost at 480p, 720p and 1080p.
Written by Sume