Showreel Studio Help
Help · Voice

Choose a voice

Pick the voice that reads a film, hear it first, and give it a direction for how it should sound.

Every film is read by a voice. The film's kit sets the voice and a few alternatives, and you can choose another for any film. The voice runs on your organisation's own Google Cloud.

Before you start

Hear a voice first

  1. Open Kits in the rail.
  2. On a kit's card, choose one of its voices from the menu and click Audition voice for a one-line read. The reads you make stay on the card to play again.

In the terminal, hear a kit read in another voice:

showreel preview voice northlight --voice Kore

Voice a film

In the app

  1. Open the film and choose the Voice panel.
  2. Choose from Voice: the kit's voice first, then its alternatives, then any own voices in your organisation, marked (own voice).
  3. Set Takes: how many reads to generate (3 by default).
  4. Click Voice the film.

When it finishes, the panel says which take is in the film, in which voice, and at what pace. See Voice takes.

In the terminal

Voice the film in the kit's voice:

showreel voice compose

Choose another voice and the number of takes:

showreel voice compose --voice Kore --takes 4

How the voice is read

The whole script is read in one continuous take, so it sounds like one person talking, then split back into lines at the pauses between them. A longer script is read in parts, broken at its chapter cards.

Give the voice a direction

A direction says how the voice should sound, in plain words, such as "warm and relaxed, for people who don't work in tech". The kit has one. A film can have its own in its storyboard file, with its own voice:

"voice": "Kore",
"voiceDirection": "warm and relaxed, for people who don't work in tech"

Write the direction as instructions to the reader, such as "Read it warmly, and keep it moving", not as sentences that could be lines of the film. A sentence that sounds like script can end up read aloud at the start of a take; when that happens, the take doesn't split into lines, and it is read again.

The two engines

The kit chooses the engine: gemini (the default) or cloud, Google Cloud Text-to-Speech. --engine picks one for a single read:

showreel voice compose --engine cloud

What next