Prosody
Also known as: intonation
Prosody is the rhythm, stress, and intonation of speech: the rise and fall of pitch, where the pauses land, and how a sentence slows into its final word. It is the performance layer, and it is usually what people mean when a voice sounds flat.
Why it matters
Prosody is what makes a voice sound like it is thinking rather than reciting. When people call an AI voice robotic, they almost never mean it is hard to understand; they mean the melody is missing. Every sentence lands with the same shape, the pauses fall on a timer instead of on meaning, and the ear stops believing a person is there.
Because it is the hardest part to fake, prosody is the real dividing line between a demo that impresses for two sentences and narration you can publish.
In practice
- Write for the ear: vary sentence length so the read has somewhere to breathe.
- Direct the lines that carry weight rather than leaving them to a default delivery.
- Judge a line inside its paragraph. A sentence that sounds fine alone can fall flat in context.
How Vocast handles this
Vocast scores prosody as a first-class metric and gates every render on it, so a take cannot pass by sounding clear while reading on a flat line.