NOTE: This table was moved to a spreadsheet in Google Sheets for easier collaboration. https://docs.google.com/spreadsheets/d/1MGCyBBHYRRBGm2kxloI5PbtkuVODWxZDioFY01zBj_g/edit#gid=579799171. The finalized results from this working document should be used to populate this page. In the meantime, please refer to the Google spreadsheet for the latest progress on modeling data properties in HydraDAM.
Concepts
- These RDF properties will be saved on the FileSet object in Fedora
- WGBH uses FITS, EXIF, and PBCore, IU uses FFProbe and has local generated info based on files received from Memnon
- ExampleTechMD.zip (from WGBH and IU)
- Additional PBCore examples from WGBH (see attached files)
- Find common properties and add those first as collaboration
- Unless properties are clearly common, they should be considered different (do not try to force fields from 2 different technical metadata profiles into a single RDF property)
- Identify appropriate predicate namespaces for these properties
- Later, expand on each of these types of technical metadata outputs to allow more detailed RDF properties to be added by expecting FITS-specific, EXIF-specific, PBCore-specific, or FFProbe-specific fields
- Align with Technical Metadata Application Profile from Hydra Metadata Interest Group, if possible
- Review against hydra-works schema characterization for base, audio, and video
Common properties
Namespaces:
ebucore: http://www.ebu.ch/metadata/ontologies/ebucore/ebucore#
premis: http://www.loc.gov/premis/rdf/v1#
nfo: http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#
Common Property | RDF Namespace:Property | SOURCED FROM | WGBH FITS | OR | WGBH EXIF | OR | WGBH PBCore | AND | IU FFProbe | OR | IU local generated info | Notes |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| File Name | ebucore:filename | /fits/fileinfo/filename | /rdf:RDF/rdf:Description/System:FileName | /pbcoreDescriptionDocument/pbcoreInstantiation/instantiationIdentifier | /ffprobe/format/@filename | PBCore: RDF mapping says instantiationIdentifer should be filename but no examples area actually reflecting that. instantiationAnnotation can contain file name but it doesn't always exist. | ||||||
| File Size | ebucore:fileSize | /fits/fileinfo/size | /rdf:RDF/rdf:Description/System:FileSize | /pbcoreDescriptionDocument/pbcoreInstantiation/instantiationFileSize | /ffprobe/format/@size | EXIF showing different value type (number with "MB") than FITS and FFProbe (just an integer, looks like bytes) | ||||||
| Duration | ebucore:duration (hydra-works using nfo:duration but I'm not sure why if ebucore can cover this property) | /fits/metadata/video/track/duration OR | /rdf:RDF/rdf:Description/Composite:Duration | /pbcoreDescriptionDocument/pbcoreInstantiation/instantiationDuration | /ffprobe/format/@duration | FITS only contains duration for video and sometimes has multiple tracks (with audio) | ||||||
| Bit Rate | ebucore:bitRate | /fits/metadata/video/track/bitRate | /pbcoreDescriptionDocument/pbcoreInstantiation/instantiationDataRate | /ffprobe/format/@bit_rate | FITS only contains bitrate for video with multiple tracks (not for audio and not for video only) | |||||||
| Format Name | premis:hasFormatName | /fits/identification/identity/@format | /rdf:RDF/rdf:Description/File:FileType | /pbcoreDescriptionDocument/pbcoreInstantiation/instantiationDigital | /ffprobe/format/@format_name | |||||||
| Sample Rate | ebucore:sampleRate | /fits/metadata/video/audioSampleRate or | /rdf:RDF/rdf:Description/AIFF:SampleRate | /ffprobe/streams/stream/@sample_rate | ||||||||
| Video Width | ebucore:width | /fits/metadata/video/imageWidth | /ffprobe/streams/stream/@width | video only | ||||||||
| Video Height | ebucore:height | /fits/metadata/video/imageHeight | /ffprobe/streams/stream/@height | video only | ||||||||
| MD5 Checksum | nfo:hashValue (Hydra Tech MAP recommends including the hash value and not the md5 checksum - it's optional. hydra-works seems to be using the nfo property for checksum so let's go with that.) | /fits/fileinfo/md5checksum | /rdf:RDF/rdf:Description/System:md5checksum | /pbcoreDescriptionDocument/pbcoreInstantiation/instantiationAnnotation/@annotationType="MD5 Checksum" | from manifest-md5.txt for each file | |||||||
| Date Created | ebucore:dateCreated | /fits/metadata/video/dateCreated | from MDPI_<barcode>.xml: /IU/Carrier/Parts/Part/Ingest/Date
|
Notes
- Fields that should be included but we don't seem to have info from both IU and WGBH:
- date created/date modified - recommended by Hydra Technical MAP; date modified is optional
- WGBH doesn't really seem to have date created but date modified info in WGBH tech md is available
- IU has date ingested from manifest but no date modified info
- can date info available be used as date created on both sides?
- UPDATE
- WGBH has date created in FITS for video from MediaInfo - not sure if audio will also have this info or what that XML path looks like
- IU has date in one manifest example and not the other, so unsure of consistency
- mimetype - recommended by Hydra Technical MAP
- IU doesn't have this - only file format (wav) and long form file format (WAV / WAVE (Waveform Audio))
- frame rate for video - used by hydra-works for video and audio schemas (not quite sure we really have to have it for starters)
- IU doesn't have this info
- date created/date modified - recommended by Hydra Technical MAP; date modified is optional
4 Comments
Michael Muraszko
Hello,
WGBH date created/date modified
Date created information is expressed in FITS version 0.8.10 forward as part of the MediaInfo data.
Julie Hardesty
Thanks! Can you pass along an example of that so I can add that to the example .zip? I can also add date created as a column with paths from both sides.
Michael Muraszko
It's in your packaged examples already, "fits_0_8_10.xml"
Thanks!
Julie Hardesty
Yeah, I missed it, thanks! It looks like WGBH might have different XML paths for date depending if the item is audio or video. And IU might not always have date - one example has it in the manifest and the other does not.