digitizing images

1
Digitizing Images Stephen Chapman Weissman
Preservation Center Harvard University
Library 30 September 2003
2
(No Transcript)
3
(No Transcript)
4
(No Transcript)
5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
Motivations
Necessity More and more material is being
produced in digital form more and more of our
users want access to such materials.
Excellence Because we value the highest levels
of teaching and research we will have to change
our way of doing things. Innovation Technology
enables uses not possible with analog formats.
Obligations are to create coherent, integrated
collections and to deliver them with tools that
support innovative professional practice.
13
Why images? The special opportunities presented
by digital technologies constitute the most
fundamental development in the potential for
increased access and flexibility of use since the
advent of photographic reproduction. Neil L.
Rudenstine, April 2001
14
(No Transcript)
15
Digital technologies
World Wide Web This is what the Internet does
well. Relational databases Permit flexible
approaches to cataloging hierarchical
structures needed to manage and describe multiple
versions of a particular item, including
surrogates. Digital cameras and scanners Proven
capability to create digital surrogates that
faithfully represent the originals in tone and
color and provide a level of detail that would
enable advanced study. David Remington
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
Products of image digitization
Images The data for scholars to study. Single
version of image rarely suffices to meet all
needs (compare, study, print). Descriptions of
images (text) The metadata the user needs to
locate and interpret images. Descriptions of
ownership and rights (text) The metadata the
owner uses to disclose terms and conditions
associated with using images. Source
materials The principal assets valued by owners
and users.
20
Infrastructure
Catalogs Systems for comprehensive and
controlled searching. Persistent naming Means to
ensure image management and reliable access.
Repository A trustworthy place to manage images
over time. Delivery Systems to deliver digital
images to authorized users.
21
Workflow
project
source
surrogates
22
Project planning There are no absolute rules for
creating good collections, objects or metadata.
Every project is unique and each has its own
goals. The key to a successful project is not to
follow any particular path, but to think
strategically and make wise choices. IMLS
Framework of Guidance
23
Selection For the kind of pictures we collect,
individual public domain analysis is expensive.
There is only one practical methodology
accession policy must choose the most
conservative boundary as a functional bright line
that separates what is acceptable and what is not
acceptable. Robert A. Baron
24
Prep Whenever there is handling of original
collections, there is a need for the application
of conservation knowledge and practice. Library
of Congress NDLP and Conservation Division
25
Workflow
project
source
surrogates
26
Cataloging practices
Local, but Picture catalogs still tend to be
incomplete, idiosyncratic, and isolated.
Helena Zinkham Movement to consolidate union
catalogs AMICO, ArtSTOR, VIA, UCAI (UCSD,
ArtSTOR, Harvard) Emerging consensus and best
practice VRA Guide, Cataloguing Cultural
Objects http//www.vraweb.org/CCOweb/ Data
standards promote sharing, improve the
management of content, and reduce redundancy of
effort.
27
Descriptive metadata standards
Specific to topics or disciplines Biology or
art Specific to kinds of materials Moving
pictures, encoded texts Specific to support
particular functions Discovery, rights
management, presentation
28
Descriptive metadata standards
Which information pieces Data dictionaries
(e.g., for OLIVIA) CDWA, VRA Core, Dublin
Core How information is formed Content standards
and vocabularies VIA Working Group has
identified over 20 How information is encoded for
processing Syntax (e.g., MARC, RDF) Virtually no
standards govern all of these aspects of
metadata. http//hul.harvard.edu/ois/systems/via/v
ia_standards.html
29
Key decisions
Scope Which catalog(s)? HOLLIS, OASIS, VIA
used at Harvard Item- or group-level
cataloging Extent Amount of cataloging (project
and program policies)
30
Digital image production

• Lights, camera
• Visual literacy and technical skill still
  absolutely critical
• Pixels!
• The more one looks at image quality and ways to
  clearly
• define it, the more parameters have to be taken
  into account.
• Frey and Reilly
• - rendering intent
• - tone reproduction
• - detail and edge reproduction
• - color reproduction
• - noise

31
Digital image standards
Formats DLF Global Digital Format
Registry Quality I3A/IT10 Electronic Still
Picture Imaging Committee ISO speed, resolution
(MTF), OECF, noise and color measurement ISO
3664 2000 Viewing conditions Technical metadata
for digital still images NISO Z39.87-2002 AIIM
20-2002 (governed by LC)
32
(No Transcript)
33
Digital imaging practices
masters
delivery images
quality control
admin metadata
support intended current and likely future
use (IMLS Framework) archival masters (optimized
for processing, not viewing) production masters
(optimized for automation) no compression for
grayscale and color images TIFF format of choice
34
Digital imaging practices
masters
delivery images
quality control
admin metadata
calibrated devices calibrated environment targets
checksums validation software at repository
35
Digital imaging practices
masters
delivery images
quality control
admin metadata

• supports management of resources (R. Wendler)
• ownership
• access restrictions
• technical attributes of files
• XML format
• produced and deposited in addition to images

36
Digital imaging practices
masters
delivery images
quality control
admin metadata
calibrated devices calibrated environment targets
checksums validation software at repository
37
Deposit DRS preservation services provide active
oversight to ensure an indefinite lifespan for
objects deposited in approved formats.
"Oversight" involves monitoring file formats,
assessing the vulnerability of digital
collections, and transforming files to maintain
usability. HUL DRS Policy Guide
38
Current cost gap digital 53-157X more expensive
than film _at_ OCLC 18-52X more expensive at Harvard
(DRS)
39
Current cost gap digital 209X more expensive
than film _at_ OCLC, 70X more expensive at Harvard
(DRS)
40
Closing cost gaps for repository storage
Compression Investigate risks associated with
using bit-for-bit lossless compression instead of
uncompressed formats as preservation
masters. Cost metrics Bill owners at unit other
than size (e.g., per GB) to sustain costs of
running repository and preservation services.
Subsidies Create common-good repositories and
services (safe havens) with secure, sustainable
funding lines for items that meet defined
criteria.
41
Hybrid approach viable for still images
Digital Masters Deposit digital master to
repository, pay for annual maintenance regardless
of use. Repurpose digital masters produce
delivery images, in analog or digital formats, in
advance and/or upon request. Analog
Masters Deposit analog (e.g., film) master to
repository, pay for annual maintenance regardless
of use. Repurpose analog masters produce
delivery images, in analog or digital formats, in
advance and/or upon request.
42
Lessons learned Building ArtSTOR into a trusted
repository will require not only time and
resources, but also collegiality and the active
participation of individuals from academic
institutions, museums, libraries, and research
centers specialists in imaging and in building
databases others experienced in the creation of
digital resources experts in intellectual
property rights and wise generalists. One clear
conclusion is that working on this project
inspires humility! William G. Bowen,
President Andrew W. Mellon Foundation
43
Resources

• Your colleagues!
• Mellon Foundation, 2001 Presidents Report,
  ArtSTOR
• Harvard University Library, LDI Program Origins
• David Remington, HCL-DIG General Imaging
  Practice
• Helena Zinkham,Bridges Whirlpools Best Access
  Practices for Pictures

44
(No Transcript)