Technical metadata - it’s all valuable

by Andrew Martin


When you decide to undertake the digitisation of your collection, you want to consider not only the importance of capturing the original content in a record that is accurate and of preservation quality, but also the associated technical metadata that is critical for the transparency of the encoding process and interrogation of information associated with the original recording.


Information obtained through the video reformatting process can include;


  • Video data (luminance and chrominance values)
  • Audio gain original quality of the signal reproduced off tape
  • Video content serial number of the playback device
  • Descriptive information such as the operator who conducted the migration work
  • All programmable settings for the migration project


Information obtained through the film reformatting process can include;


  • Perforation detection including bad perforation
  • Film speed zoom and focus settings of the camera
  • All colour grade values


For preservation purposes it is ideal that all information is encoded as XML, which is desirable because it’s an open and flexible standard. Metadata can be stored as a separate XML file, or within the video wrapper incorporated into the video file.  


Why store all that metadataI hear you ask? For a 200min duration video you can end up with an xml file of around 100MB. For the sake of the argument, if the video was encoded using a standard definition uncompressed codec, it would end up around 300,000 MB, so when comparing the XML file size to its much larger associated video file, it puts the amount of metadata into a relevant perspective.


How can it be utilised?


  • As a document of the migration process, auditable and accountable
  • A report for use in collection assessment for example, assessing the amount of monochrome content in a collection, or determining the original video quality of a particular format
  • Scene detection, and edit decision lists for non-linear editing systems
  • Offline interrogation of video (If large preservation size files are offline in a HSM environment, or the video content is restricted and stored in a dark archive)
  • Motion detection for marking scene changes
  • Uploading technical information into for Media Asset Management Systems
  • Determining physical quality of the carrier


We have been using XML reports in ways we hadn’t considered when we started collecting the technical metadata for our clients. In one example we’re parsing or interrogating timecode information to determine accuracy of the audio head reader in the VTR. In another we’re injecting new timecode into the XML, and then transferring the new information into its associated video file. We also use in-house software tools to parse information from the XML as a step in our quality control procedures.


Bottom line?


Retain all technical information because it is extremely useful now, and we don’t know how useful it could be in the future.