35 min
Lesson 3 of 5
AI Voice Cloning & Overdub
From the course Descript: AI Video & Podcast Editing
Descript's Overdub feature creates an AI clone of your voice that can speak any text you type.
Creating Your Voice Clone
- Record training data: Read provided scripts for 10-30 minutes
- Submit for processing: Descript trains a model on your voice
- Verify ownership: Consent and identity verification required
- Voice ready: Use your AI voice in any Descript project
Using Overdub
Once your voice is trained:
Fix Mistakes
- Find the error in your transcript
- Delete the incorrect words
- Type the correct words
- Descript generates the correction in your AI voice
- The correction blends seamlessly with surrounding audio
Add New Content
- Place cursor where you want to insert new audio
- Type the new text
- Overdub generates speech in your voice
- No need to re-record entire sections
Create Narration
- Write a script in the Descript editor
- Select your Overdub voice
- Generate complete narration without recording
- Adjust pacing and emphasis
Stock AI Voices
If you don't want to clone your own voice:
- Library of professional AI voices
- Multiple languages, accents, and styles
- Suitable for demos, prototypes, and internal content
- No voice training required
Best Practices
- Train with quality audio: Use a quiet room and good microphone
- Read naturally: Don't over-enunciate during training
- Provide variety: Training scripts include different emotions and pacing
- Review outputs: Always listen to Overdub content before publishing
- Ethical use: Only clone your own voice or voices with explicit permission
Key Takeaways
- •Overdub creates an AI clone of your voice from 10-30 minutes of training recordings
- •Fix recording mistakes by typing corrections that generate in your voice seamlessly
- •Complete narration can be created by typing scripts without any voice recording
- •Voice cloning requires consent and identity verification for ethical use
0 of 5 lessons marked complete
Progress is kept in this browser only.