Skip to main content

Characters · Lesson 7 of 8

How to make a talking avatar from a photo in any language

Green belt· Intermediate0 of 6 done toward your Green belt1:50 videoUses P-Video Avatar

P-Video Avatar turns one portrait and one typed line into a talking video, with 30 voices in 10 languages and a price per second.

You need: A clear, front-facing portrait and the line they should say.

Step by step

  1. Add a portrait

    Open P-Video Avatar and click the plus to upload a portrait or pick one from your library. A close-up of a saved character works best.

  2. Type the line

    Hi, I'm Juniper. Give me one photo and one line, pick a voice, and I'll say it in any of ten languages. Ready? Let's make something.

    The length is estimated from your script.

  3. Pick a voice and a language

    Open the voice menu: 30 named voices. Pick the language beside it. Have a recording already? Press Add audio and it lip-syncs to that instead.

  4. Pick the quality

    You pay per second: 0.5 credits at 480p, 1 at 768p, 1.5 at 1080p, up to 3 at 4K. A 10-second line at 1080p is 16.5 credits, shown on Generate.

  5. Generate and save

    Press Generate. About a minute later they say every word in sync. Press Save to keep it.

Why it works

  • Write the line the way people talk: short sentences give the most natural delivery.
  • Same face, new market: generate the same line in another language with the same portrait.

Made this way

Juniper says her line at 1080p: one portrait, one typed sentence, 16.5 credits.

Your practice task

Make the same 10-second line in two languages with one portrait.

Read the transcript(1:50)

Turn one portrait into a talking avatar in PixelDojo with your own script, a chosen voice, and adjustable video quality. Start in P-Video Avatar. This page brings together your portrait, script, voice, and quality settings with 30 voices across 10 languages available for your talking video. Add your portrait first. Here, Juniper's photo is already attached above the text box. That portrait supplies the face for the talking avatar you're making. Next, enter the script exactly as you want it spoken. The duration estimate below the text box comes from those words, so check it after typing. Open the Voice menu beneath your script. It currently shows Zephyr. Choose Kore for this example, giving Juniper the voice we'll use for the generated line. Open the Quality menu to compare resolutions. Pricing is per second, from half a credit at 480p to 3 credits at 4k. For this video, choose 1080p, listed as Full HD Upscale. Its rate is 1.5 credits per second between the lower and higher resolution options. With 1080p selected, this 10-second line comes to 16.5 credits. The total appears right on the button. Check it, then press Generate. The generation card now says, Starting, confirming the request is underway. Give it about a minute, and Juniper says every word with her mouth movements in sync. Hi, I'm Juniper. Give me one photo and one line, pick a voice, and I'll say it in any of 10 languages. Ready? Let's make something. You've turned a portrait and script into speech. Next, try another voice with the same portrait in P-Video Avatar.

Go deeper