Voice name check in synthesize is too strict and missing short name support

jhudsl / mscstts

R Client for the Microsoft Cognitive Services Text to Speech REST API

GNU General Public License v3.0

8 stars 6 forks source link

Voice name check in synthesize is too strict and missing short name support #10

Closed boltomli closed 4 years ago

boltomli commented 4 years ago

Short name is added to be usable in official request. The check is too strict and limit the usage now.

Also, if the specified voice is not in embedded voice list (which is not region aware and may not align with the actual voice list from endpoint), synthesize will fail early. It could be a warning if failing this check and continue. Then if it's really not correct, service can return error.

boltomli commented 4 years ago

This is related to https://github.com/muschellij2/text2speech/issues/2. Currently I have to mix text2speech and mscstts, aws.polly together in https://github.com/boltomli/shiny-speech.

boltomli commented 4 years ago

With #11, short name can be used as voice, though the actual voice name attribute is still in long format in the actual payload. It's easy to change the behaviour as well but I would like to keep it as it is now for minimal change.