Hi guys! I want to create app for voice messenger or metaverse room where different nation people can speak in their own voices in bad english but application will correct them pronunciations and they speech will be clear. What can I do?
Thanks for your interest in Riva
I will check regarding this use-case with the internal team and let you know the updates
I have inputs from the internal team
if the model is finetuned on your data which has multiple English dialects, then the model will be robust and will work fine for people speaking different english accents.
In Riva ASR , there is no autocorrect as such but there is alignment. And how good the generated transcripts are , depends upon the ASR model finetuned on various dialects.
Riva ASR can be used with the finetuned ASR model (on English dialects) and you can make an app for your use case.
Sample Apps for reference