Columbia University Libraries Digital Program

METS Metadata Requirements: Description

The vendor will supply a set of METS files for each interview.

1. For every Interview, there is one high level METS file containing:

1. A dmdSec containing Dublin Core descriptive metadata for the interview
   
   
2. A sourceMD section with a pointer to an external "audio decision list" (adl) file which reports which sections of the source audio were used to construct the rendered audio
   
   
3. A structMap broken into 2 divs: one for the source audio, one for the rendered audio
   
   1. source audio contains one div for each tape which has any content related to the interview at hand, with an mptr to a METS file fore each digitized tape
      
      
   2. rendered audio contains an mptr to a single METS file for the audio content drawn from the source audio which has been edited into continuous uninterrupted audio

2. For each digitized source audio object, there is a single METS file containing:

1. A dmdSec containing Dublin Core descriptive metadata for the source audio object
   
   
2. A techMD section containing the JHOVE XML output from a JHOVE process run by the vendor for each wav file produced
   
   
3. A techMD section containing AES Core Audio metadata for each wavfile produced
   
   
4. A fileSec with file elements for each wav file produced by the vendor
   
   
5. a structMap containing divs for each "face" (usually tape side) of the source audio.

3. For each rendered audio object there is a METS file containing:

1. A dmdSec containing Dublin Core descriptive metadata for theentire interview and each interview session
   
   
2. A techMD section containing the JHOVE XML output from a JHOVE process run by the vendor for each wav file produced
   
   
3. A techMD section containing AES Core Audio metadata for each wav file produced (i.e, each session or session part)
   
   
4. A sourceMD section containing a pointer to the ADL file mapping the portions of the source audio which were used to create each portion of the rendered audio
   
   
5. fileSec's for the hi-res ("rendered") audio and low-res ("service") wav files (one each for each session or session part)
   
   
6. A logical structMap breaking down the interview into sessions