Speed comes from the sample
A cloned voice copies how its sample was spoken, including how fast. A sample read quickly gives quick takes; a slow, relaxed one gives slow takes. That makes the sample the single strongest control over pace: if every line comes out too fast, re-record the sample at the speed you want, and replace the voice's clip.
Three settings that affect pace
- Guidance strength (in Fine-tune) is how closely the take follows the clip's pacing. If your sample is fast and the takes rush, lowering it to around 0.3 loosens that grip and usually gives a steadier, more deliberate read.
- Emotion intensity: turning it up tends to speed speech up along with the drama. If you raise it, lower guidance strength a little to compensate.
- The Calm style (lower emotion, higher guidance, steadier expressiveness) suits measured narration; Expressive loosens everything.
The voice settings guide explains every slider.
Pauses come from punctuation
Before a line is read, its punctuation is tidied into a few marks the model knows well. Knowing what becomes what saves a lot of re-takes:
| You write | The voice gets | So expect |
|---|---|---|
Full stop . | A full stop | A clear pause and a falling ending |
Comma , | A comma | A short breath |
Ellipsis ... or … | A comma | A short pause, not a long trailing-off |
Colon : or semicolon ; | A comma | A short pause |
Spaced dash - | A comma | A short pause |
| Question or exclamation mark | Kept | A rising or lifted ending |
| Extra spaces | One space | No pause at all |
| No punctuation at the end | A full stop is added | A normal ending |
So the reliable tools are full stops and commas. More full stops means more, clearer pauses; splitting one long sentence into two short ones slows a read down more than any setting.
Line breaks are the exact control
Every line in an abcVoiceover script is its own take and its own audio file. The silence between two lines isn't decided by the voice at all: it's whatever gap you leave when you place the files in your editor. For a dramatic beat, a pause before a punchline, or a long rest in a meditation, end the line where the pause goes and start a new one. See editing in Audacity for spacing lines.
Long lines
A line can be up to 2,000 characters. Long ones are made a few sentences at a time, about 20 seconds of speech per piece, and joined with a short gap about the length of a sentence break. That keeps a long paragraph steady, but it does mean you don't control the pauses inside it as precisely. If a long line has a pause in the wrong place, split it into two lines.
Small speed changes in the edit
If a finished take is just slightly too fast or slow, audio editors can change its tempo without changing the pitch, such as Audacity's Effect › Pitch and Tempo › Change Tempo. Keep it within about 5 to 10 percent; beyond that it starts to sound processed. Bigger changes are better made with a new sample or a rewritten line.
Quick fixes
- Rushing: lower guidance strength, add full stops, or record a slower sample.
- Dragging: raise guidance strength slightly, cut filler words, or record a livelier sample.
- No pause where you want one: use a full stop rather than a dash or an ellipsis, or split the line.
- A pause in the wrong place: remove the comma, or rephrase so the sentence doesn't break there.