enhancementhelp wantedspoken-to-signed
Repository-Metriken
- Stars
- (777 Sterne)
- PR-Merge-Metriken
- (Durchschn. Merge 1m) (1 gemergte PR in 30 T)
Beschreibung
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