I have to say, after looking at the text to speech model on VP23, I am very disappointed on how robotic and unreal some of them sound. I mean, its not even like its new technology anymore. I think you guys should either remove it or up your game and do some rework on it.
Sorry to be negative, but I just wanted to be honest.
I think you guys should either remove it or up your game and do some rework on it.
Text-to-Speech (and v-v) is, I believe, sourced from Microsoft (possibly MS Azure) and, yes, some of the voices are quite robotic but some - depending on the language/accent - provide a choice of emotions (e.g. happy, sad) for the dialogue read.
Earlier today, I watched a YT video on the history of 20th century movie lots in California. It clearly used a US accented AI text-to-speech read and was so bad that I stopped watching halfway through. And it wasn't helped by the creator not allowing 'breath' room or a pause of any kind between sentences, paragraphs or chapters.
I understand that there are 3rd party AI text-to-speech providers that have been suggested for Premiere Pro and Davinci Resolve users as neither of those NLEs natively feature text-to-speech (and v-v) - unless that has recently changed with Premiere Pro. But then you'd need to be sure of the source of voices for those 3rd party plugins to make sure that it's not also MS.