Which Live Stream Workflow Is Best for Live and Hybrid Events?

Introduction:

For live and hybrid events, you need to cater to both the on-site audience and potentially several online audiences. This often involves displaying transcripts or translations on a screen at the event while simultaneously streaming a video with multiple subtitles online.

To keep the process low cost and easy to manage, it’s important to avoid complex setups like live editing of subtitles or splitting video signals for redundancy. Instead, a streamlined workflow can meet both requirements efficiently.

Live Stream Setup for On-Site and Online Audiences:

Typically, the video stream from the live event is sent to Deep Live Hub via RTMP. On the output side, a text stream with transcripts and translations can be sent to an auto-generated website for on-site or mobile consumption. Additionally, the stream can be pushed to a video platform or content delivery network (CDN) via RTMP, with either burned-in subtitles or closed captions.

Burned-In Subtitles:

• Only one subtitle language can be burned in per workflow. For multiple subtitle languages, you need to split the stream and send parallel workflows, one for each language.

Closed Captions:

• Closed captions allow up to four subtitle languages embedded in the stream, but accessibility to these languages depends on the video platform or player in use.

Time Sensitivity at Live and Hybrid Events:

Live events often require minimal latency as the on-site audience needs to follow the speaker’s actions in real time. Online audiences may also need to interact with the content, for example, by participating in a poll.

To reduce latency, consider setting the ASR (Automatic Speech Recognition) configuration to 3-5 seconds in partials mode. This will allow the on-site audience to follow the transcripts with a delay of less than 1 second. However, translations will still be delayed by the ASR processing time plus an additional 2 seconds for translations. For online audiences, burned-in subtitles add another 8 seconds of delay.

Creating a Live Stream Workflow for Webcasting:

1. Login to Deep Live Hub:

   • Click on the red "+ Workflow" button and select "Live Stream."

2. Choose a Workflow Template:

   • Select the template for the platform you’re using: "YouTube," "Twitch," or "Custom" for other platforms.
   • Enter a name for your workflow and click "Finish."

3. Configure the Workflow:

   • Navigate to the Process tab and choose your translation languages from the dropdown menu.
   • If you have dictionaries, select them to improve recognition of specific terms.
   • Choose an ASR configuration with a 3-5 second processing time in partials mode for low latency.
   • Click "Save" to confirm your settings.

1. Set Up Outputs:

   • In the Outputs tab, you can configure different outputs such as:
     • SRT Configuration: Generates separate SRT files for transcripts and translations.
     • Video Output Pull Configuration: Adds subtitles to the video stream if burned-in or closed captions are enabled.
     • Text Configuration: Use partials mode for near real-time transcript updates.

1. Push to a Video Platform:

   • Create a Video Output Push Configuration to send the stream to a CDN or video platform. Enter the URL and credentials, then click "Save."

2. Collect URLs:

   • In the Overview tab, collect the URLs for input (RTMP, WebRTC) and your stream key.
   • You’ll find the egress URLs (Latest SRT, RTMP, HLS, HLS-realtime) and a link to the Live Viewer for transcripts.

3. Managing Streams:

   • View past and current streams under the Streams tab.
   • To disable or delete a workflow, go to the Settings tab. Stopping the stream in your encoder first will deactivate the workflow.