AI GitHub
Chatterbox for controllable text-to-speech
Generate speech and experiment with reference audio and documented expression controls.
Key Takeaways
- The project offers speech synthesis models and example inference code.
- Select a documented model variant and its required voice assets.
- Use an authorized reference recording and compare a short neutral line with a more expressive delivery.
- Reference quality and model variant affect the result.
How it works
The project offers speech synthesis models and example inference code. Depending on the variant, a voice reference or control settings can influence the output.
What you need
Select a documented model variant and its required voice assets. Use short consented recordings and explicit text so differences in tone and pronunciation can be judged separately.
What to try first
Use an authorized reference recording and compare a short neutral line with a more expressive delivery. Listen for skipped words and inconsistent voice quality.
Try two different reference recordings from the same consenting speaker. If outputs vary substantially, the recording conditions may matter more than the control setting you changed.
Where to look for failures
Reference quality and model variant affect the result. Use consented recordings and verify pronunciation before distributing audio.
Related Franklin pages
Use authorized voice references and compare both intelligibility and similarity. Check the exact checkpoint terms before moving from a local test to published audio.
- Explore voice cloning with a reference speaker and supported synthesis components. OpenVoice for experiments in voice tone conversion
- Experiment with published generative audio models and training components. AudioCraft for music and sound generation research
For an application where the quality of recording and handling audio matters, see Franklin’s Open Mic AI profile. Open Mic AI
Project documentation
Try it for your workflow
Explore the project’s documentation for setup and usage, or save this page to a collection for later.
Read the project documentation ↗