Prosody "presets" (Volume, Pitch, Rate) matched to theatrical voice descriptors, including terms like these:
- grumble
- ambivalent
- harsh
- sarcastic
- scornful
- arrogant
- impassioned
- indignant
- inquisitive
- sincere
- jovial
- solemn
- thoughtful
- elated
- threatening
Here's a little background on the project and rationale on Medium
<prosody rate="fast" pitch="high" volume="loud">This voice is very excited!</prosody>
SSML can be tested using IBM's interactive demo
Any help would be much appreciated! I'm no expert in the human voice but this library may prove extremely useful to anyone developing a VUI tool.
Simply add an xml document to the repository containing the prosody tag and attributes to emulate a style of voice. Make sure to name the document with your term. For example the above lives in excited.xml.