To create a broadcast websocket you need to be in Live Transcribe Mode. Broadcast Websocket management is under Live Broadcasting menu.
Creating a new Websocket
To start creating a new Broadcast Websocket click on the Add Broadcast Websocket button. This will bring up the following dialog:
You will need to enter:
- The name of the websocket which has to be globally unique - easy way to create a globally unique websocket is to prefix it with your company name, e.g. voicegain-meetings-dev
- The context which will determine:
- the default settings for the speech recognizer (ASR settings)
- the security settings (JWT for authentication)
- context for transcript output if using `portal` output destination (this applies only if using Web API directly)
Once a websocket is created we can click the (i) icon/button to see the details of its configuration and modify it. This will bring a page like the one shown below.
What you see here is mixture of websocket settings together with controls for the Audio-Sender Daemon.
The websocket settings are:
- context - already explained above
- Private/Public setting - private websockets will not be shown on the public company page on the Voicegain Web Console
- Default/Non-Default setting - only one Websocket per account may be default. I will be implied when trying to reference websocket without name. (this specific setting is mainly of importance for Live Transcription on the Edge/OnPrem)
- Delay - this is for the secondary time-delayed feed - might be useful is some scenarios to sync to delayed streaming video feed, e.g., and even broadcast on Vimeo. (Note: this may be removed in the future as this feature is almost never used).
Controlling the Audio Daemon
We also have controls and info relevant to the Audio Sender Daemon which sends the audio to the speech-to-text engine.
- Status -one of: "Offline", "Online", "Online-Listening", "Online-Muted"
- play/stop button - controls the Audio Daemon - switches be "Online" and "Online-Listening" states
- disabled if Audio Daemon is "Offline"
- mute/unmute buttons - switches between the "Online-Listening" and "Online-Muted" states
- disabled if Audio Daemon is not "Online*"
- download the Audio Sender Daemon tar file.
- you need to choose the language model first.