Version 22.1

22.1

2022-03-01

Closed caption extraction and burn-in

The workflow of extracting closed caption (cc) data from video and burn-in cc into the lowres file has been improved in this release (#197489).

The cc data will be extracted as item metadata if the extractCc parameter is set to true on placeholder/raw import jobs. And the extracted cc data will be burned into the proxy file if a shape tag with burnSubtitles=true

<TranscodePresetDocument>
  <video>
    ...
    <burnSubtitles>true</burnSubtitles>
  </video>
</TranscodePresetDocument>

For example, to extract and burn-in cc at the same time during an import job, use:

POST /import?uri=...&extractCc=true&tag=burnsubtitle-tag

Be able to add custom notification metadata on delete endpoints

It’s now possible to attach custom metadata to item, shape or file deletion requests using the notificationMetadata parameter. For example:

DELETE /storage/file/{file-id}?notificationMetadata=deletion-reason=workflow1

The custom metadata will be included in the corresponding entity deletion notification.

New features

  • Be able to set framerate on image sequence import (#197468)
  • Add item or collection in a collection using external-Id (#197335)

Improvements

  • Send s3 uri with roleArn directly to Vidinet (#200658)
  • Bulkymetadata migration - delete bigtext entry directly (#199510)
  • Bulkymetadata migration - Balancing batchsize better (#199053)
  • Switch to use bytea instead of large object for postgresql (#198875)
  • Attach user name to the notification (#197355)

Bug fixes

  • Fix incorrect AUTO-VSA baseuri if there is a port conflict while starting VSA noauth service (#201523)
  • Can not import FCP projects with 10 fps clips (#201278)
  • startTimecode from TranscodePreset isn’t copied to NLE output (#200667)
  • Keep a margin to EOF while reading growing files (#200355)
  • Fix incorrect hostname in AUTO-VSA url (#200205)
  • Some ACLs are lost after item been removed from a collection (#199502)
  • Bulkymetadata on storage - Deletion too slow, can cause ItemDeleteCruncher to time out (#199149)
  • Bulkymetadata migration - File access issues can kead to data loss (#199045)

Transcoder fixes

  • Support for extracting timed text subtitles (#200249)
  • Support for specifiying framerate when transcoding image sequences
  • Support for setting start timecode of NLE jobs

Upgrading from 21.4

  1. Database migration is needed to migrate all columns of oid type to bytea type.
  2. Solr: No changes to the documents. Re-indexing is not required.