Our interface currently only allows for single turn translation, where for example a user can input text, then we perform the translation of the entire text.
This would allow continuous speech, and using the browser's native sentence segmentation we could translate one sentence at a time and stitch the sentences together in the client
Problem
Our interface currently only allows for single turn translation, where for example a user can input text, then we perform the translation of the entire text.
Description
For spoken-to-signed, we can allow a "continuous mode", perhaps looking like this https://codepen.io/abnersn/pen/ExVwoWB
This would allow continuous speech, and using the browser's native sentence segmentation we could translate one sentence at a time and stitch the sentences together in the client
Alternatives
Wait until everything is spoken, then translate.
Additional context
No response