Designing a Voice
Designing a voice is how you create a reusable target voice for later audio workflows.
Designing a Voice
Designing a voice is how you create a reusable target voice for later audio workflows.
Once saved, that voice can be used in:
- text-to-speech generation
- voice change / speech-to-speech conversion
What the designer does
The voice designer does not create one final voice immediately. It first generates preview options. You listen to the previews, choose the one you want, and then save that preview as your reusable voice.
How to open it
The main entry point is the Voices folder in Asset Manager:
- Open Asset Manager
- Go to Voices
- Click Create Voice
You may also reach the same designer from audio workflows that need a saved voice first.
What you set up
The voice designer lets you control things such as:
- Voice name
- Voice description
- Model
- optional reference audio
- Guidance
- Prompt strength
- Quality
- Loudness
- Enhance audio
- optional seed
Voice description requirements
The voice description needs to be meaningful enough for the system to work from. Very short descriptions are not enough.
A useful description usually focuses on:
- age impression
- tone
- energy
- clarity
- accent or language feel
- delivery style
How to design a voice
- Open Asset Manager
- Go to Voices
- Click Create Voice
- Enter a clear voice name
- Write the voice description
- Optionally drag a reference audio clip into the reference slot
- Adjust generation settings if needed
- Generate the previews
- Listen to the returned preview voices
- Select the preview you want
- Save it as your final voice
Using reference audio well
Reference audio is optional, but it is useful when you want stronger guidance for tone, pacing, or delivery style.
It helps with:
- overall tone
- delivery style
- pacing feel
- vocal energy
Reference audio is guidance, not a perfect one-to-one clone.
What happens after saving
Once saved, the voice appears in your available Voices list and can be selected later in text-to-speech and voice change workflows.
What to do next
After designing a voice, the usual next step is one of these:
- create speech from text in Creating Text-to-Speech Audio via Asset Manager
- record a performance in Recording Voice-Overs
- convert existing audio in Changing a Voice-Over into a Target Voice