View a markdown version of this page

Setting up Smart Subtitles (console) - MediaLive

Setting up Smart Subtitles (console)

Follow these steps to enable Smart Subtitles in a channel using the MediaLive console.

  1. On the Create channel page, ensure that you have added at least one input attachment to the channel. An input attachment is required before you can add a Smart Subtitles caption selector.

    If you want Smart Subtitles to transcribe a specific audio track, add an audio selector to the input attachment before proceeding to the Elemental Inference settings.

  2. Navigate to the Elemental Inference settings section.

  3. For Elemental Inference feed, select a feed that contains a subtitling output. If you need to create a new feed, choose Create feed at the top of the Elemental Inference settings section. If you need to update an existing feed, choose Create or update the feed in Elemental Inference beneath the Subtitling output dropdown. For more information, see Creating an Elemental Inference workflow in the AWS Elemental Inference user guide. After creating or updating the feed, choose Refresh to see the updated feed list.

    If the input attachment has audio selectors, for Audio selector name, select the audio selector that contains the audio you want to transcribe.

  4. Under Smart Subtitles, choose Add Smart Subtitles caption selector. This adds a caption selector to the input attachments.

  5. Configure the caption selector:

    • Caption selector name – Enter a name for the caption selector (for example, SmartSubtitlesSelector1).

    • Language code – Choose the language of the audio in your source media. This must match the language configured on the subtitling output of the selected feed.

    • Caption synchronization mode – Choose how MediaLive synchronizes generated subtitles with the video output:

      • Video aligned captions (default) – MediaLive delays video to ensure captions are synchronized with audio and video.

      • No video delay – MediaLive does not delay video for caption alignment. Caption output timing is adjusted to align with video as captions become available.

    • Subtitling output – Select a subtitling output from the selected feed. If the feed does not have a subtitling output, update the feed in the Elemental Inference console to add one, then choose Refresh to update the list of outputs.

  6. Create a captions output for the subtitles:

    1. In your output group (HLS, MediaPackage, MediaPackage V2, CMAF Ingest, or Microsoft Smooth), add a new output.

    2. Remove any video and audio encodes from this output so that it contains only captions.

    3. Add a captions description and set the destination format to TTML (for MediaPackage V2, CMAF Ingest, or Microsoft Smooth output groups) or WebVTT (for HLS or MediaPackage output groups).

    4. For Caption selector name, select the Smart Subtitles caption selector you created earlier.

Note

Before configuring Smart Subtitles, you must have an Elemental Inference feed with a subtitling output. You can create the feed in one of the following ways: