I had done some experiments with their SSML, expressive SSML, and Voice Transformation SSML in 2017. It was an interesting way to change, almost like coding the voice in order to make it more human-like. Went back to IBM Watch Text to Speech demo today: https://www.ibm.com/demos/live/tts-demo/self-service/home, and found out it works differently now.
“Hurry up! Pick up the sword and defend yourself, your arrival has awakened the spores. It’s not a good thing.”
“Greetings, my name is Luke, the voice that you’re hearing right now is not my real voice. It is a result of my telepathic thoughts being synthesized by your auditory cortex. It could take form of any voices you have heard before.”
“incoming”