Talking Head

Portrait to video

Create a talking head avatar from a portrait

A talking head avatar turns a still face into a speaking video. Start with a clear, front-facing image and a short script, then review the result for natural mouth movement and a recognizable likeness.

Review your output before sharing
Talking Head portrait video example

The input portrait and script matter more than elaborate instructions. Gather both before you begin.

One full run-through

For a first attempt, make one short greeting rather than a long presentation. That makes timing, pronunciation, and facial movement easier to check.

  1. 1

    Prepare the portrait

    Pick a sharp, front-facing photo with the full face visible. Crop out distracting surroundings, but leave enough space around the head that movement does not feel cramped.

  2. 2

    Write the spoken line

    Draft one or two sentences in the language you want the avatar to speak. Read them aloud and replace awkward names or abbreviations with wording that is easy to pronounce.

  3. 3

    Generate and inspect

    Submit the portrait and words, then watch the entire clip with sound. Check whether the mouth follows the audio, the face stays recognizable, and the final pause feels complete.

Options table

A photo-based avatar preserves a chosen face; a generic generated speaker is better when no specific person needs to appear. Neither option replaces reviewing the finished video.

1

Starting input

Your portrait

A photo you provide and have permission to use

Generated speaker

A described or selected character

2

Identity

Your portrait

Aims to resemble the person in the image

Generated speaker

Does not need to resemble a real person

3

Best fit

Your portrait

Personal introductions and messages from a known presenter

Generated speaker

Explainers that do not require a particular presenter

4

Preparation

Your portrait

Check lighting, crop, face visibility, and image rights

Generated speaker

Choose an appropriate appearance and presentation style

5

Main risk

Your portrait

Distortion of recognizable facial features

Generated speaker

A speaker who feels too generic for the message

6

Final review

Your portrait

Check likeness, lip movement, and consent

Generated speaker

Check character consistency, speech, and suitability

What fails

A talking head clip can communicate a script, but it cannot repair every weak input or guarantee a convincing performance.

  • A hidden or angled mouth

    A hand, microphone, heavy shadow, or extreme profile leaves too little facial detail for reliable lip movement.

    WorkaroundUse a brighter, more front-facing portrait with the mouth unobstructed.

  • Unnatural pacing

    A long sentence with no natural breaks can sound rushed even when every word is spoken correctly.

    WorkaroundSplit the script into shorter sentences and add punctuation where the speaker should pause.

  • Guaranteed likeness

    Animation may alter small features or expressions. A recognizable starting photo does not guarantee that every frame will look like the person.

    WorkaroundWatch the full result and try a different source image if the face drifts.

  • Permission by default

    Having access to someone's photo does not mean you have consent to make a video that appears to speak for them.

    WorkaroundUse your own image or obtain clear permission before creating or sharing the clip.

Try a short portrait message

Give your portrait one clear line to say

Start with a well-lit image and a brief script. Generate a draft, then check the face, voice, and wording before you share it.

Animate my portrait
  • Use a photo you have permission to animate
  • Keep the first script short
  • Review the complete video with sound

Talking head avatar FAQ

It is a video presenter whose face appears to speak on screen. It may be animated from a supplied portrait or created as a generated character.

A clear photo of your face is a suitable starting point for a portrait-based workflow. Use an image with even lighting and a visible mouth, then review the video for likeness and lip synchronization.

That depends on the workflow you choose. Some workflows use written text to produce speech, while others use supplied audio; check the available input options before preparing your script.

An obscured face, difficult pronunciation, or rapid speech can make the movement look less convincing. Try a clearer portrait and shorter, more naturally punctuated lines, then watch the new result from beginning to end.

Only use a person's likeness when you have permission to do so. Be especially careful with videos that could imply the person said or endorsed something they did not.

Start creating
Start creating