I’m not sure if anyone has used the speech API much yet, so this question may be one that only @simeon can answer.
With the speech API is there a way to print on screen the word that is spoken? For example, if I have a table of words and one is randomly chosen to be spoken, can that randomly chosen word be printed on screen?
Surely it’s a simple case of writing a function that speaks the word and prints it on screen?
Hey, yeah I realized that I was way overthinking it. Thanks!
There is a more in-depth API I am hoping to expose relating to this, I would like to have a callback available to tell you when the speech system is about to read a specific portion of text. This would allow you to highlight the exact part of the text (e.g., syllable) as it is being spoken. Kind of like Karaoke.
@Simeon would it be possible to expose the phoneme that is currently being uttered? i.e. via some ascii-based system of transcribing phonetics, like the pronunciation guides you see in some dictionaries? I don’t know if the back-end api supports this. I’m thinking for facial animation, determining mouth shapes. You could try to do it with the word portion, but English spelling being as un-phonetic as it is, a phonetic stream would be more accurate.
@yojimbo2000 you’d have to map the portion of the string about to be spoken to the phoneme. Though considering the synthesised voice is fairly monotonous, I don’t think it would be too great paired with facial animation.
Would it be possible to set the language? Right now the langage is automatically the local language, which is not practical for reading an english text in Paris…