This containes code to simulate emotional arousal with SSML tags.
A paper with evaluation is
Going Retro: Astonishingly Simple Yet Effective Rule-based Prosody Modelling for Speech Synthesis Simulating Emotion Dimensions
The code is a running flask http application and can be found here
Here are some demo audio files
| arousal/valence | 0.0 | 0.5 | 1.0 |
|---|---|---|---|
| 0.0 | wav | wav | wav |
| 0.5 | wav | wav | wav |
| 1.0 | wav | wav | wav |