Audio Technical Metadata

From UA Libraries Digital Services Planning and Documentation
Jump to: navigation, search

For our WAV files, we currently are generating and storing output from the File Information Tool Set (FITS), as well as AES57 (Audio Engineering Society standard for audio metadata - Audio object structures for preservation and restoration).

During quality control testing, the QC script used by staff members generate FITS files and tests them for errors. Staff members review the errors, and if warranted, repair the file and test again.

The scripts which generate derivatives for the web test the upload directories to ensure there's a FITS file for each WAV, and will not continue until there is.

The scripts which upload the WAVs for deposits to be processed for the archive, retest the FITS files for errors. If none are found, this script also generates AES57 files and uploads some information to our md5sum:audioTechMed database table for quick reference. (More information on this database below).

A free version of the file is available at fits2aes.

In the archive, the FITS and AES files are named for the WAV, with extensions ".aes.xml" and ".fits.xml". If the file was repaired or modified, there will be versions of these files as well: the first version (fileNumber.aes.v1.xml and fileNumber.fits.v1.xml) would be from the first version of this file; v2 from the second version, etc. The unversioned AES and FITS files are the CURRENT ones. All technical metadata is stored in a Metadata directory at the level to which it applies. So if this is technical metadata for a page, then in the archive on the page level there will be a Metadata directory containing the technical (and any other) metadata for that page.

Example: /srv/archive/u0008/0000581/0000003/0002/Metadata/ will contain metadata about page 2 of the 3rd item in the WC Gorgas collection (Collection MS 581) in the Hoole library audio holdings:

  • The tiff is named u0008_0000581_0000003_0002.wav
  • The current MIX file is named u0008_0000581_0000003_0002.aes.xml
  • The current FITS file is named u0008_0000581_0000003_0002.fits.xml

The md5sum database.

The information currently stored in the md5sums:audioTechMed table includes:

  1. the identifier,
  2. mime type,
  3. format,
  4. format version,
  5. date of testing,
  6. format registry,
  7. format registry key,
  8. whether there are any conflicts between the findings of the various FITS tools,
  9. and if any conflicts were found in the identity of the file:
    1. what the alternative format,
    2. format version,
    3. and mime type are.