Audio Generation
Generate speech from text with fine-grained emotion control. This tool uses voice cloning — provide a reference audio sample and the AI generates speech in that voice.
How to Use
- Type the text you want spoken
- Upload a reference audio — A sample of the voice to clone (max 20 seconds)
- Choose an emotion control method
- Click Generate
Emotion Control
Four methods are available for controlling the emotional tone of the generated speech:
| Method | Description |
|---|---|
| Same as Reference | Matches the emotion of your reference audio |
| Use Emotion Reference | Upload a separate audio file whose emotion will be transferred |
| Use Emotion Vector | Manually set emotion values (Joy, Anger, Sadness, Fear, Disgust, Gloom, Surprise, Calm) |
| Use Emotion Description | Describe the emotion in text |
Emotion Vector
When using the emotion vector, you can set values for 8 emotion dimensions. Each value ranges from 0 to 1, and the total sum should not exceed 1.5:
- Joy, Anger, Sadness, Fear, Disgust, Gloom, Surprise, Calm
Requirements
- Reference Audio — A clean voice recording, max 20 seconds duration
- Emotion Reference Audio — Optional, max 20 seconds duration
- Text — The content to be spoken
Computing Power
Audio generation is free (0 CP).