Is it possible to add some disfluency words like "mmh-mh", "huh", "uh", "erm", "um"," "hmm" into a Dialogflow conversation (only interacting by voice) using SSML? Currently, when I add "mmh" to a sentence, it will be being spoken as characters m-m-h but not as a disfluency word.
Any ideas?
On
I tried to test disfluency words pronunciation using TTS simulator in Action console (select your project > click on Test > click on Audio) and most of these words sound well. I tested "uh", "oh", "ah", "erm", "hmm", "um", "aha" and all of them sound normal, but some of them like "mhm", "mmh-mh", "mmh" spelled as characters. You need to create Feature Request (select "Create new Cloud Text-to-Speech issue") and add what words or interjection should be added to the library.
Have you tried using Oxford English words that sound like disfluencies?
For example, "em" (as in em dash) might sound like ummmm. "Ohm" and "home" might also sound like ummmm.