Support SSML
compulim opened this issue · 0 comments
compulim commented
Today, we wrap the text as a SSML.
Tomorrow, we should detect if the text looks like SSML. If yes, we should not wrap it and pass it to Cognitive Services as-is.
The user should not specify pitch/rate/volume/voice while they are passing SSML to SpeechSynthesisUtterance. Otherwise, they will be ignored.