Prosody

The <prosody> tag enables fine-tuning of voice characteristics such as volume, pitch, rate, and intonation, allowing for nuanced vocal responses from conversational AI bots.

Settings Scope

• Local Settings – Found in all blocks which include SSML tags.

• Global Settings – Located in Releases, and are accessible during the addition or editing of a release.

The Prosody settings, including global and local configurations, are only available for the Phone platform.

When turned ON, global settings are applied across all responses of the bot unless specifically overridden by local settings within individual blocks. If local settings specify 'rate', 'pitch', and/or 'volume' (but not 'contour'), they will take precedence over global settings.

For instance, using the <prosody> tag could make the bot speak with increased speed, volume, or pitch to convey urgency. Conversely, for a more relaxed and slower response, you can fine-tune the prosody settings as needed.

Attributes of the Prosody Tag

Contour

Specifies pitch variation or pattern for the text:

• linear: Pitch rises or falls linearly.

• exponential: Pitch rises or falls exponentially.

• none: Maintains a steady pitch.

• default: Uses the default setting of the speech synthesis engine.

Pitch

Specifies the pitch variation or pattern for the text.

• Relative: Sets a relative pitch (x-low, low, medium, high, x-high). Default pitch is medium.

• Semitones: Adjusts pitch by +/- N semitones (+Nst or -Nst).

• Percentage: Modifies pitch by +/- N% (+N% or -N%) relative to the default pitch.

Note: % is required, +/- is optional.

Rate

Modifies the speaking rate of the text:

• Custom – Expressed as a multiplier of the default rate (for example, 1 means no change, 0.5 halves the rate).

• Predefined –

  • x-slow: Very slow speaking rate.

  • slow: Slow speaking rate.

  • medium: Normal speaking rate.

  • fast: Fast speaking rate.

  • x-fast: Very fast speaking rate.

  • default:

• Percentage –

  • N%: A speaking rate that is N% faster than the default speaking rate.

  • -N%: A speaking rate that is N% slower than the default speaking rate.