Skip to main content
Skip to content
Generate speech from text

Generate speech from text

Choose a voice and turn written dialogue into speech.

Before you start

Open a storyboard shot, or in the editor choose Add Media, then Audio, then Generate speech. Speech is on every plan. The script can be up to 5,000 characters. The charge is 1 credit per 10 characters, rounded up. The Generate speech button shows the total.

Steps

  1. On a shot, choose Speech & Sound FX, then Speech, then Open speech controls. The window title is Generate speech. In the editor, Add Media → Audio → Generate speech opens the same window. The editor subtitle is Add a voice line to your project. On a shot it is Give this shot a voice.

  2. Choose Text to speech. If no voice is selected, the window says No voice selected and the button is Choose a voice. Choose Change voice when one is already selected. The picker title is Change speech voice. Choose Use on a voice. Create voice, on this tab, opens design and clone.

    Generate speech in the video editor, on Text to speech, with Rachel from ElevenLabs, Speech script, Choose from script, cue chips, and Voice stability set to Natural.
  3. Enter the line under Speech script, or choose Choose from script. The field holds up to 5,000 characters. Add a cue offers [laughs], [whispers], and [excited].

  4. Under Voice stability, choose Creative, Natural, or Robust. Creative varies the delivery. Robust keeps it more consistent. Natural is the starting choice. Fish Audio voices can show Pace and Volume instead.

  5. Choose Generate speech. The button shows the credit total: 1 credit per 10 characters, rounded up. After a take exists, the button can say Generate again. In the editor, choose Add to project to keep the take.

Expected result

The recording is on the shot, or in the editor after you choose Add to project. It is also saved with your speech files. Place it on the timeline when you want to control when it plays.

Tips

Changing the voice does not rewrite a recording you already generated. Generate again makes a new take. Speech to speech is a separate tab and needs Plus or above.

Documentation