When will new voices support blendshape output?

Question

Hello,

we are using the text-to-speech service and are relying on blendshapes for facial animations. However, some voices do not support blendshapes and this doesn't seem to be documented. In the voices overview (https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=tts) there are footnotes when visemes are not supported, but this doesn't help in our case as there are voices, which support visemeIds, but not blendshapes. Also, https://<region>.tts.speech.microsoft.com/cognitiveservices/voices/list does not have any fields indicating whether blendshapes are supported or not.

Am I missing where this is documented?

And is there a plan to support blendshapes also for voices where they are missing at the moment?

Best regards,
Steffen

Answer 1

Hello @VasaviLankipalle-MSFT
Is there anythign new regarding this issue?
We noticed that not only do some voices not have blendshapes, but blendshapes seem to be actively removed from some voices as well. Compared to last week, for en-US locale, Brian and Emma do not support blendshapes anymore, but did so, last week. Can you explain this? I was hoping that blendshapes support would rather increase than decrease.
I would really appreciate to have some way of getting noticed about such changes before finding out ourselves that some voices do not work anymore in our setup.
Best regards,
Steffen