Make a Hindi or Urdu Talking Avatar with Midjourney and D-ID

Turn a Midjourney portrait into a talking avatar in D-ID Studio. Use a short Hindi or Urdu script and preview the pronunciation, or upload your own recording for control over delivery.

Midjourney V8.2 is the current default. Midjourney’s web and Discord tools require a subscription, and D-ID Studio’s trial has plan limits and a watermark. Midjourney trial availability · D-ID Studio plans.

Create and animate the avatar

  1. Create a portrait with a clearly visible face. In Midjourney Create, use V8.2 to generate a front-facing, well-lit portrait. Choose a neutral expression with a closed, unobstructed mouth and save the image. A front-facing cartoon character with distinct lips generally gives you a better starting point for lip sync.

    Original friendly 3D cartoon presenter, front-facing head-and-shoulders portrait, neutral expression, lips clearly visible, soft even lighting, plain blue background --v 8.2 --ar 1:1

    Full Midjourney Create page with the Imagine bar and settings outlined
    Use the Imagine bar to enter the prompt and its settings icon to check the model.
  2. Add the portrait in D-ID Studio. In Studio, select Create video. Open the Avatar controls and upload your portrait. Keep the face centered and large enough to be readable in the finished frame. Use a portrait you created or have permission to animate.

    Full official D-ID Studio screenshot with Avatar controls highlighted
    Open Avatar to choose or replace the presenter, then select Script.
  3. Add Hindi or Urdu speech and preview it. Open Script. In Text, enter your script, choose a voice for the language if available, and use the play control to check pacing and names. If a suitable Urdu voice isn’t available or pronunciation matters, use Audio to record or upload speech you have permission to use. Record in a quiet room without music. Test a sentence before creating the full video.

    Official D-ID Script panel with Text and Audio highlighted
    Use Text for a synthetic voice or Audio for your own recording.
  4. Generate a short test, check it, then download. Select Generate video and review the credit cost before confirming. Watch the completed clip for lip sync and audible speech. If the mouth movement is unclear, use a more front-facing image or rerecord the audio. In the gallery, open the finished video’s three-dot menu and select Download.

    Official D-ID editor with Generate video highlighted
    Review your script and avatar before selecting Generate video.

D-ID’s guide to image requirements, audio formats and downloading.

What still works from the original tutorial

The 2023 workflow creates a character with /imagine in Discord, saves a favorite from the grid, and imports it through the older D-ID presenter picker. Its most useful test is to try a few faces and choose the one with the clearest front-facing mouth. The original script and audio tabs also show the same two approaches available today.

Full Discord screen showing a Midjourney portrait grid and its selection buttons
Compare all four portraits and select one whose eyes and mouth are clearly visible.
Full D-ID editor with the presenter upload, script area, and voice selector outlined
Upload the portrait at the left. Add the script or audio and choose or preview a voice at the right.

Use the preview to check the mouth movement before committing to a longer script. The finished example below shows the lip-sync result.

Full preview of the generated talking avatar with the face and mouth outlined
Watch the generated preview to check the mouth movement before downloading.

Original Discord and D-ID tutorial

Share:

0 comments:

Post a Comment

Search This Blog