Federal Digitization Moving to Common Guidelines The U'S' Federal Agencies Digitization Initiative h

1
Federal DigitizationMoving to Common
GuidelinesThe U.S. Federal Agencies
Digitization Initiativehttp//www.digitizationgu
idelines.gov/

• NDIIPP Partners Meeting, June 2009
• Carl Fleischhauer
• cfle_at_loc.gov
• Michael Stelmach
• mste_at_loc.gov
• Library of Congress
• Washington, DC

2

• http//www.digitizationguidelines.gov/

3
Participating agencies . . .
4
http//www.digitizationguidelines.gov/stillimages/
5
Advisory Board
6
http//www.digitizationguidelines.gov/audio-visual
/
7
(No Transcript)
8
(No Transcript)
9
Selected use case objectives for master images

• Digitizing organization (or successor/ receiving
  agency with an archiving mission) sustains the
  master (or migrated copies) for the long-term
  without loss of essential features.

10
Selected use case objectives for master images

• Digitizing organization uses master to produce
  derivative images for use cases like these
• (1) end-user-access interface
• (2) other patron uses as listed
• (3) OCR or other text-creation process
• (4) document the condition of the original item

11
Selected use case objectives for derivative
(service) images

• Publisher uses image to illustrate a book.
• Publisher uses image to illustrate a large
  poster.
• Exhibit designer uses image for display "mural."
• Broadcaster uses image in high-definition
  television program, zooming in for Ken Burns
  effect.

12
Selected use case objectives for derivative
(service) images

• Patron sees inline image or image set in
  interface. Some view the complete work, a virtual
  replica.
• Patron prints images. Some require
  print-on-demand copy of complete work, a physical
  replica.
• Patron is confident that the content received is
  an authentic reproduction, also receives
  information on restrictions.
• Patron downloads a derivative image and, later,
  uses embedded metadata to identify content and
  determined technical provenance.

13
Plan to move from specifications with these
factors color/monochromatic pixel density
(good old dpi) bit depth . . . usually
output-referred
To specifications with these factors
14
Working document from the National Library of the
Netherlands. Three columns, three categories.
Specifications in the various rows.
15
Tools to Support Image
Performance Measurement

• Digital Image Conformance
  Evaluation (DICE) System
• Device Target Imaging Device Performance
• Object Target Actual Image Quality
• Software for Evaluation/Validation
• Based in LabVIEW
• Data export for use in SQC/SPC

16
Device and Object Targets
Object target as positioned for use Thanks to
OCLC for help with this part of the effort.
17
DICE Software Main Panel
18
DICE QC Summary Panel
19
(No Transcript)
20
Beyond performance measurement

• Embedding metadata
• TIFF header specification online now
• Future exploration of XMP

21
Beyond performance measurement

• Other gaps in prior guidelines to be
  investigated
• Image Sharpening
• Quality Management
• Image Specification Metric Aims and Limits
• Foldouts and Inserts in Bound Materials
• Color Encoding Accuracy
• Color Space Encoding
• Selection Criteria for Master Image File Format

22
Working draft pertaining to quality assurance and
quality control
Work in progress at the National Archives and
Records Administration
23
Audio-visual effort recorded sound

• Compile guidelines for recorded sound

Work in progress
24
(No Transcript)
25
Audio-visual effort recorded sound
26
Audio-visual effort video

• While we wait for agencies to gain experience . .
  .
• Exploration of target formats

27
Library of CongressPackard Campus, Culpeper
National Archives, College Park
Smithsonian Institution Archives
28
Lossless compressed

• Each frame is a JPEG 2000 image
• Lossless (reversible) transform
• Produced by the SAMMA device

29
(No Transcript)
30
What about film?

• Most activity is service to outside customers,
  usually television documentary makers
• Addressed by making a video copy, often still
  standard definition, understood to be an
  imperfect solution

31
Most active high-resolution film scanning
program NASA Johnson Space Center
32
Please review our work and pass along your
comments http//www.digitizationguidelines.
gov/contact/
33
(No Transcript)
34
One of the subcategories

• T.3. Documents with poor legibility or diffuse
  characters, e.g., carbon copies,
  Thermofax/Verifax, etc. manuscripts or
  printed/typed pages with handwritten annotations
  or other markings items with low inherent
  contrast, staining, fading, printed halftone
  illustrations, or included photographs.

35
One of the subcategories

• Valuation determined
• by curator or end users
• to have informational
• and artifactual value,
• but not requiring color
• reproduction.

36
From this document http//www.digitizationguidel
ines.gov/stillimages/documents/Digital_Imaging_Fra
mework.pdf
37
Image recommendation in 2004 guidelines
from NARA

• 8-bit grayscale mode - adjust scan resolution to
  produce a QI of 8 for smallest significant
  character
• or
• 8-bit grayscale mode - 400 ppi for documents with
  smallest significant character of 1.0 mm or
  larger
• NOTE Regardless of approach used, adjust scan
  resolution to produce a minimum pixel measurement
  across the long dimension of 4,000 lines for
  8-bit files

38
Uncompressed video

• Stanford, Rutgers
• 422 or 444, 10-bit SDI stream
• About 100 GB per content-hour
• Another source reported 70 GB for 8-bit video

Rutgers spec http//rucore.libraries.rutgers.edu/
collab/ref/dos_avwg_video_obj_standard.pdf
39
The Netherlands Institute for Sound and Vision
Lossy compressed
MPEG-2 _at_ 50 mbps
and 30 mbps (news)
40
SONY IMX, MPEG-2 _at_ 50 mbps
From http//www.edithouse.com.au/information/imx.
html

• MPEG-2, all I-frames, 50 mbps
• File size about 28 GB/hour
• MPEG-4 (ITU-T H.263 and H.264) may come to play
  a bigger role as high-resolution increases

41
Lossless compressed
42
Audio-visual effort video

• Video reformatting target format
• Federal Agencies Working Group planned action
  Documentation of MXF wrapping JPEG 2000 and
  uncompressed video

43
Emerging encoding preferences

• For high value, uncompressed or lossless
  compressed is very attractive.
• For second-rank content, some make a good case
  for modest-but-lossy compressed.

44
(No Transcript)
45
Audio-visual effort recorded sound

• System performance testing
• Considering IASA TC04 pass-fail specifications
• Appropriate, affordable equipment for tone
  generation not at hand

46
Audio-visual effort video

• Target encoding options
• Uncompressed
• Lossy compressed
• Lossless compressed
• File wrapper options
• MXF
• AVI, QuickTime, other