You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Description:
We are using the Azure OpenAI Realtime API to build a conversational avatar for live user interaction. The avatar is expected to pause or stop immediately when the user starts speaking, as part of a natural conversation experience. However, the current behavior continues speaking over the user's input, creating a robotic and frustrating experience.
Expected Behavior:
When the user starts speaking while the avatar is talking, the avatar should immediately stop and yield control to listen to the new input.
Current Behavior:
The avatar continues speaking even when interrupted by user audio.
There is no mechanism for dynamically stopping speech or pausing mid-sentence upon user input.
This leads to overlapping speech, making it feel robotic and unnatural.