What an AI talking head video generator does
A talking head video is the simplest format in video: one person, framed from the chest up, speaking to the camera. It is used for explainers, announcements, course intros and the opening lines of ads. Filming one normally needs a presenter, a quiet room and a few takes.
An AI talking head video generator removes the filming. In AdsTurbo Lip Sync, a still portrait supplies the face and a voiceover supplies the speech; the AI generates the lip movement from the audio and returns a video of that face delivering your lines. The portrait does not have to be a real person. Non-human characters are supported as long as the face and mouth are clearly visible.
What to prepare: the face and the voice
- Portrait: one clear, front-facing image with the face and mouth visible, as a JPG or PNG up to 10 MB. WebP and GIF are not supported.
- Framing: the video keeps the aspect ratio of the portrait, so crop it to 9:16, 1:1, 16:9 or whichever shape the video needs before you upload.
- Voice, option one: upload a recording as MP3, WAV, M4A or AAC, up to 10 minutes. Keeping the file within 50 MB makes the upload more stable.
- Voice, option two: type a script of up to 500 characters and generate the voiceover with text-to-speech in the workspace.
Photos with several people, a side-facing subject or a partly hidden face tend to give weaker results. Background music under the voice can also reduce quality, so use a clean voice track when you can.
How to create a talking head video in AdsTurbo
Add the portrait
Open Lip Sync and upload the portrait, or choose an available source image.
Add the voiceover
Upload your audio file, or enter a script and generate the voice with text-to-speech.
Check the estimate and submit
The workspace shows the credit estimate for the length of your audio before you submit the task.
Review the result
The finished 1080p video appears on the Project page. Check lip timing, audio clarity and how the face looks before you use it.
Writing for a talking head
The delivery is decided by the audio, not by a setting in the tool. Tone, pace, pauses and emphasis all come from the voiceover, so get them right before you upload. If you record your own voice, read the script aloud once to find the spots where you run out of breath, then shorten them.
Length drives cost. Video generation currently uses 24 credits per second of audio, so a 20-second talking head is 480 credits and a one-minute explainer is 1,440. Creating the voice with text-to-speech does not use credits, and credits are returned if generation fails. For ads, several short versions with different opening lines on the same portrait usually teach you more than one long take.
Where it fits, and where it does not
- Good fit: product explainers, ad openers, onboarding or course intros, and the same message recorded in several languages with one portrait.
- It starts from a still image. You cannot upload an existing video of a presenter and re-voice it here; AdsTurbo Video Translation handles dubbing of existing videos.
- Audio is not translated automatically. Prepare a separate voiceover for each language.
- Captions are not included in the output. Add them afterwards with AdsTurbo Video Subtitle or your editor.
- Use portraits and voices you have the rights to, and get consent before using a real person’s face or voice.
Frequently asked questions
How do I make a talking head video with AI?
Open Lip Sync in AdsTurbo, add a front-facing portrait as JPG or PNG, upload a voiceover or generate one from a script of up to 500 characters, then submit the task. The 1080p talking head video appears on the Project page.
Do I need to record my own voice?
No. You can upload a recording as MP3, WAV, M4A or AAC up to 10 minutes, or type a script of up to 500 characters and generate the voiceover with text-to-speech in the workspace.
How long can a talking head video be?
The video follows the length of the audio. Uploaded audio can be up to 10 minutes, and a text-to-speech script can be up to 500 characters.
How are credits calculated?
Video generation currently uses 24 credits per second of audio. Text-to-speech does not use credits, and credits are returned when generation fails. Check the estimate in the workspace before submitting.
Can I turn an existing video into a talking head with new audio?
No. This workflow starts from a portrait image. To translate or dub a video you already have, use AdsTurbo Video Translation.
