Skip to main content
Center active speaker is an AI Tools action that runs through Underlord, Descript’s AI editing assistant.
On current plans, this feature uses AI credits.

Reframe your video

Use video that includes its recorded audio, including when it is part of a multi-track sequence. Select your aspect ratio, and Underlord will create scenes and crops for each active speaker.
  1. Open your composition.
  2. Choose Center active speaker from AI Tools > Look good in the AI Tools panel.
  3. Select an aspect ratio: Landscape (16:9), Square (1:1), or Portrait (9:16).
  4. (Optional) Choose your preferred model.
  5. Click Generate to kick off the task.
Processing can take several minutes; Underlord will notify you once the task is complete. After running center active speaker, you can edit the video layers in each scene.

Known limitations

  • This feature is in beta. Review the output before exporting.
  • Center active speaker works best with a single wide-angle video of multiple speakers. If you’re working with multiple speakers that have dedicated camera tracks, we recommend using Automatic Multicam instead.
  • Crops stay fixed within each scene. They don’t continuously track a moving speaker.
  • Footage with existing camera cuts can produce late or incorrect framing. Check speaker changes carefully, including in sequences, and adjust affected crops.

Related topics

Automatic Multicam