← All posts

How to Frame a Portrait for a Popstar AI Video

October 15, 2026 · Popstar AI

For a Popstar AI video, start with one sharp, front-facing portrait that shows the face and shoulders in even light. The template turns that single image into a fixed 12-second performance exported in a 9:16 vertical format; choosing a clean source photo matters more than writing instructions because the motion is preset.

Framing choice Recommended Avoid
Subject count One obvious subject Multiple competing faces
Face angle Front or slightly turned (about 30° or less) Profile or sharply turned head
Crop Head and shoulders with breathing room Cut-off head or tiny distant face
Light Even exposure across the face Deep shadow or strong backlight

Start with one clearly visible subject

The Popstar AI workflow takes one portrait photo as its input. Choose an image with one obvious subject, with the face large enough to distinguish and the shoulders visible. A distant group shot can make it unclear which person should become the lead.

The photo supplies the identity for the template. It does not provide extra frames or let you choose a different performance, background, or camera move. If you want a particular person to appear in the result, make sure that person is the clear subject of the uploaded image and has consented to its use. This guidance is about human portraits; do not assume non-photographic or pet images will work reliably.

Keep the face close to front-facing

A straight-on or slightly turned portrait gives a clear view of facial features. Avoid profile shots and strong head turns, which hide part of the face. The site specifically advises against angles beyond roughly 30 degrees, as well as hands or other objects covering the face.

Check the eyes, brow, nose, mouth, jaw, and hairline before uploading. A fringe across the eyes, a hand under the chin, or sunglasses can hide details that may matter to maintaining a recognizable subject during motion.

Use even light and avoid heavy filters

Choose a well-lit photo without a hard shadow crossing the face. Backlighting can make the face too dark, while colored or uneven light can obscure its outline. Heavy beauty filters may soften or reshape the brow and jaw, so a natural, unfiltered image is a safer starting point.

The original file should also be sharp enough to see facial details. If the face is blurry when you preview the image, try a clearer portrait rather than expecting the render to recover detail that is not present in the source.

Leave room around the head and shoulders

Avoid cropping tightly at the top of the head or at the shoulders. A little space around the subject gives the source image a more complete outline. At the same time, do not use a full-body image where the face becomes a tiny part of the frame.

The generated video uses the template's vertical 9:16 composition. The portrait itself does not let you reposition the camera or choose a custom aspect ratio. For that reason, a clear head-and-shoulders photo is a practical input, while the final framing remains controlled by the template.

Understand the fixed video format

Popstar AI produces a 12-second clip with the template's motion and original audio. The workflow does not include a free-text prompt field for changing the walk-in, scene, or timing. Selecting a better portrait can improve the input, but it cannot customize those fixed parts of the video.

The available output tiers are 480P, 720P, and 1080P. A generation costs 45 credits at 480P, 90 at 720P, or 180 at 1080P. The tiers affect output resolution, not the source image or preset performance. The current credit packs and pricing are shown on the Popstar AI pricing page; confirm the displayed options before you render.

Use this checklist before uploading

  • One person is clearly the intended subject.
  • The face is sharp, unobstructed, and nearly front-facing.
  • The head and shoulders are visible, with a little space around them.
  • Light is even, without strong shadows or heavy color effects.
  • The image is one you own or are authorized to use.

If a finished result looks warped, try a different photo that improves the specific issue you can see: a more direct angle for a partly hidden face, more even light for unclear features, or wider framing if the head is cropped. Higher resolution changes the export tier, not the underlying portrait or preset performance.

Share synthetic video responsibly

Use images only when you have the right and permission to do so. A generated likeness can still affect the person shown, even when the source photo is publicly available. When sharing the result, do not present it as real footage of an artist or another person; make clear that it is AI-generated where appropriate.

When you are ready, upload the selected portrait in the Popstar AI studio and review the available render options before spending credits. If the finished face is unclear, match the visible problem to the input: choose a more direct angle for a turned face, more even light for lost features, or a less tight crop for a cut-off head. If the portrait is already clear but the fixed scene or movement is unsuitable, another crop or resolution tier will not provide custom direction.