Adding Pauses in Voiceover Transcripts

When using the AI to generate speech, you may want to introduce pauses or breaks to control the rhythm and cadence of the speaker.

Voice Variability

The way the AI handles pauses can vary depending on the voice used. Some voices, especially those trained with a few "uh"s and "ah"s in them, may sometimes insert those vocal mannerisms during the pauses, mimicking a real speaker.

While less consistent, you can use a simple dash ("-") or em-dash ("—") to indicate a pause. Multiple dashes can be used for a longer pause.

Ellipsis ("...") can also work to add a pause between words but may convey a sense of hesitation or nervousness in the voice.

Note: Using excessive SSML breaks in your text may cause issues such as speeding up of speech, increased noise in the audio, or other artifacts. We are working on resolving this. For consistent results, we recommend using the SSML syntax for pauses.