AV Prototype Project Working
Documents
Explanation: DigiProv (Digital Provenance) Extension Schema
February 2003
Introduction
Conceptual Example
Comparisons to the AES approach.
Actual Example
Introduction. Our digiProvMD (digital provenance
metadata) elements and schema owe a great debt to work in progress under
the auspices of the Audio Engineering Society (AES), especially the contributions of David
Ackerman of the Harvard University Library Digital Initiative Group. Our prototyping has been
mostly concerned with audio, and as we began to implement digiProvMD, our
experiences led us to rearrange some elements within the emerging AES approach, and to
introduce a few simplifications that seemed possible in our particular circumstances.
One of our applications for digiProvMD concerns the physical
preparation of source materials we are reformatting, e.g., a 1/4-inch magnetic tape or a 78
rpm disc. In our METS structMap we assign one level (div) to the
original item, for example, Tape or Disc (actual wording
may vary). We associate one instance of sourceMD (metadata about the thing we
start with) and one instance of digiProvMD with that node in the structMap. For
example, if we baked the tape (to combat sticky-shed syndrome) or cleaned the disc side, then the instance digiProvMD at this node in the structMap documents the baking or cleaning of the original item.
Our second common application for digiProvMD has to do with how a specific file
was made. When we reformat a sound recording, for example, we make three digital audio files:
- Master. Our best copy in terms of preservation-quality reproduction, typically a 96
kHz/24-bit WAVE file.
- Higher-resolution service file. A very good quality copy for research listening, a 44
kHz/16-bit WAVE file.
- Lower-resolution service file. A good quality, lossy-compressed listening copy, typically
a 256 kbps MP3 file.
These three files are referenced in the structMap "below" the original
item, i.e., Disc is the parent of Side A, which is in turn the parent of three file children. The references takes the form
of fptrs (file pointers) that link back to the METS fileGrp (file group)
section and thus to the individual files. Associated with each file are instances of (1)
techMD (technical metadata) that provides information about file characteristics,
e.g., how many samples per second, what compression, and (2) digiProvMD that
documents the devices and settings used in the reproduction
process. (To be perfectly accurate, we try to always provide an instance of
digiProvMD for masters; this provision is optional for service files.)
As we have carried out our prototyping, we have seen (but not yet realized) an intimate
relationship between workflow management and the use of workflow management systems and
software to capture the data that would finally emerge in instances of digiProvMD.
This is something for the future.
Conceptual Example. The following example is illustrative and is intended
to communicate some of the concepts we have used.
- Process. Our starting point. Example: Digitize Audio Master from LP.
- Task. Child of process. In our one-task example below: Digitize Audio Master from
LP. We have had some two-task examples in the past, e.g., Copy from 78 rpm disc
to DAT tape and Make files from DAT tape.
- Tools. Child of task, typically more than one. Examples: Turntable, Phono Cartridge,
Phono Pre-amp, A-to-D Converter, DAW (digital audio workstation), etc.
- Settings. Child of tools. Any tool may have settings. For example, the Phono Pre-amp can be set to the Standard RIAA equalization for LPs.
Most of our processes are used again and again. Therefore, our data-capture software permits an
operator build a process template, e.g., Joe's lp digitizing process. In
data entry, the template is referenced by the operator in the database's instance table along with
organization and individual identifiers and today's date. Later, when the XML is output, all of
the information from the process template and the instance table is incorporated into the
digiProvMD section of the final METS document.
Comparisons to the AES approach. The following comments are not
exhaustive but describe our understanding of some relationships between our approach and that
of the AES:
| AES element group(s) |
LC variant (tables in the database used for
initial metadata capture) |
| Event |
Instance plus Task |
| Device, Module (hardware), and Channel |
Tools |
| Settings, Parameter, and Preset |
Settings |
| Module (other than hardware) and Auxiliary
File |
Configuration File (Table name: ConfigFile) |
Actual Example.
<?xml version="1.0" encoding="UTF-8"?>
<!--Sample XML file generated by XML Spy v4.4 U (http://www.xmlspy.com)-->
<DIGIPROVMD xmlns="http://www.loc.gov/PMD/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://www.loc.gov/PMD/http://lcweb2.loc.gov/mets/Schemas/PMD.xsd"
ID="ID000000" TYPE="DIGITIZE_AUDIO">
<process_creator>Joe Smith</process_creator>
<process_datetime>2002-05-16T03:57:53</process_datetime>
<process_description>This is the process that describes taking a physical source (lp in
this case) and digitizing it into the Master audio file.</process_description>
<process_label>Digitize Audio Master from LP</process_label>
<process_organization>MBRS</process_organization>
<task ID="ID000001">
<task_label>Digitize Audio Master from LP</task_label>
<task_description>This takes the physical source (lp) and converts it into a 96/48 uncompressed wave file.</task_description>
<task_sequence>1</task_sequence>
<task_individual>Joe Smith</task_individual>
<task_organization>MBRS</task_organization>
<task_datetime>2002-06-20T10:47:59</task_datetime>
<tool ID="ID000002">
<tool_label>Turntable</tool_label>
<tool_description>Simon Yorke Turntable Precision Analog Disc
Transcription System</tool_description>
<tool_make>Simon Yorke</tool_make>
<tool_version>PAT</tool_version>
<tool_serial_number>L03</tool_serial_number>
<tool_sequence>1</tool_sequence>
</tool>
<tool ID="ID000003">
<tool_label>Phono Cartridge</tool_label>
<tool_description>Phono Cartridge</tool_description>
<tool_make>Shure</tool_make>
<tool_version>V15</tool_version>
<tool_sequence>2</tool_sequence>
</tool>
<tool ID="ID000004">
<tool_label>Phono Preamp</tool_label>
<tool_description>Phono Preamp</tool_description>
<tool_make>EQS Mark12</tool_make>
<tool_sequence>3</tool_sequence>
<settings ID="ID000005">
<settings_label>Equalization</settings_label>
<settings_description>Standard RIAA equalization for
LPs</settings_description>
<settings_value>RIAA</settings_value>
</settings>
</tool>
<tool ID="ID000006">
<tool_label>A/D Convertor</tool_label>
<tool_description>Analog to Digital Convertor</tool_description>
<tool_make>Apogee</tool_make>
<tool_version>PSX100</tool_version>
<tool_sequence>4</tool_sequence>
</tool>
<tool ID="ID000007">
<tool_label>Sadie DAW</tool_label>
<tool_description>Sadie Digital Audio Workstation</tool_description>
<tool_make>Sadie</tool_make>
<tool_version>96/24 w/ Software: 3.4</tool_version>
<tool_sequence>5</tool_sequence>
</tool>
</task>
</DIGIPROVMD>
Go to top
Go to Extension Schema overview
Go to AV Prototype Project Documents
Go to AV Prototype Project Home
(2/25/03)
Library
of Congress
Comments: AV
Prototype Coordinator (cfle@loc.gov)
Legal | External Link Disclaimer
(
August 31, 2010
)