Speech Generation

RenderAI provides text-to-speech, which allows users to generate and export audio of realistic, human-sounding speech for use in audio/video content. It’s as sinple as selecting a voice, and writing a script.

Prompt Engineering

The AI automaticlly infers emotion and tone from the text. If you need more control over the delivery, you can try writing the script in the style of a book.

Additionally, punctuation and capitalization can be used to add emphasis.

For example:

Completly shocked and utterly disgusted, she finally managed to say,

“are you KIDDING ME?

I just… can’t even believe this.”

Settings

Randomness

A higher randomness setting produces a wider variety of tones and emotions over the course of the performance (more range).

Voice Fidelity

The higher this setting, the closer to the original cloned voice.

Performance Boost

The higher this setting, the more that the idiosyncracies of the voice (i.e. accent, unique inflections, patterns of speech, etc) will be exaggerated.