• Feature Article
• Imaging Pictorial Collections at the Library of Congress by John R. Stokes
• Technical Features
• DoD-NARA Scanned Images Standards Conference by Sue MacTavish
• Highlighted Web Site
• Calendar of Events
• Announcements
• FAQs
• RLG News
• Hotlinks Included in This Issue

John R. Stokes, Manager, Digital Imaging, JJT Inc.
johnr@jjt.com

Introduction
As part of its National Digital Library Program, the Library of Congress (LC) selected a range of pictorial collections and other visual media to digitize and make available on the Web. A Request for Proposals (RFP97-9), for the conversion of pictorial materials was issued, and a contract awarded on September 16, 1997 to JJT, Inc., a software development and imaging service company based in Austin, Texas. This article describes the technological and production approach taken by JJT in this project.

The contract required that very high quality digital images be produced for a variety of original pictorial materials and formats: 222,500 b/w negatives ranging in size from 2x2 to 8x10 inches; 11,600 b/w and color prints ranging in size from 1x1 to 11x14 inches; 6,500 color negatives and (non-slide) transparencies, and 5,650 color slides. The Library choose both photographic and graphic materials for scanning, including baseball cards and printed duotones by William Gottlieb. A continuous, 34-month effort was identified for the on-site scanning in the Prints and Photographs Division of the estimated 170,000 photographic negatives in the Historic American Buildings Survey/Historic American Engineering Record (HABS/HAER) collection and smaller groups of prints. Initially, a second capture site at the Library's Adams Building was to focus on small- or medium-scale projects of coherent groups of related materials, but JJT convinced the Library to centralize all in-house scanning at the first site. The contract required that a minimum of 525 images be captured each day onsite. In addition, the Mead-Bateson Bali-New Guinea photographic negatives were slated for scanning at the contractor's facility.

The MARC II Camera
After a thorough evaluation of commercially available cameras and scanners, JJT concluded that none provided the requisite quality and production capabilities needed to meet the requirements of LC's contract. JJT had already discovered from work performed for the Library under another contract that a substantial amount of information was lost in correcting the tonal distribution of negative images captured by equipment that internally translated the 12-bit input capture into an 8-bit output format. JJT had previously decided to develop its own image processing and management software, and in the end decided to build the complete digital imaging system, including custom designed digital cameras that would provide access to the "raw" data from the Analog-to-Digital Converter.

Udo and Reimar Lenz, CCD Videometrie, agreed to design and build three cameras similar to the MARC cameras they developed for the European MARC (Methodology for Arts Reproduction in Color) project several years earlier. The MARC II cameras built for JJT offer higher resolution and more sensitive sensors, near frictionless mechanical screws on the stepping motor, and more advanced electronics than the first MARC cameras. Hence, the MARC II requires only a fraction of the illumination, has significantly faster capture speeds, and is designed to ensure consistent performance over long periods of time.

As currently configured, the MARC II can capture originals at various resolutions up to 9,920 x 9,375 pixels. A 7,500 x 5,700-pixel resolution image can be scanned in grayscale (b/w) in 34 seconds and in color in 53 seconds. Image assembly and write times are effectively hidden by use of a multi-processing workstation. The images are captured at 12 bits and left shifted (multiplied by 16) to create 16 bits per pixel per color. All color and tonal adjustments and processing are performed in the 16-bit mode. The images are converted to 8 bits for b/w and 24 bits for color images immediately before writing to final format.

Overview of JJT's Work Station at the Library of Congress

Image Processing and Management Software
Keeping track of images, managing the image processing, and writing to new medium can be a labor-intensive process. JJT developed software to track the images, control the image processing, and record every event affecting the image from input to final verification of the images written to the storage medium. The image processing and management software operates under the umbrella of a master program called Convert. It includes tools for capturing, viewing, analyzing, manipulating, and processing images. The user has the option of processing individual images or batch processing. There is also an option of incorporating the image tracking database software. When operating in this mode, the database functions as a control mechanism for each process. Various other tools can be incorporated into Convert as needed.

The image tracking database is centered on an Oracle database server. C++ is used primarily for active image manipulation and preparation; SAS and SQL are used for database maintenance and reporting; and Java is used for user interface to data analysis and presentation. The software is the core of the digital imaging system. Its development was by far the most demanding, time consuming, and expensive part of the system, but it has enabled JJT to meet the quality and production goals. Refinements to the software continue to be developed.

Camera Stand, Lighting, and Peripherals
The camera stand consists of a column with motor driven camera and lens position, mounting assembly, a table with a film platen, overhead and/or bottom light sources, and a laminar airflow canopy. The laminar airflow canopy consists of a HEPA filter mounted above the camera stand with an anti- static plastic curtain attached to the filter and draped around the stand.

Stepping motors and controllers provide precision positioning of the camera and lens. Follow focus is achieved via a program within the imaging processing software. The software fine-tunes the focus by measuring the contrast between pixels and adjusts the lens accordingly.

The transmissive light source consists of a diffusion box with four 12-volt quartz halogen lamps placed about 16 inches below the diffuser. A two-inch air space separates diffuser and the platen. The reflective light source consists of two v-shaped reflectors with four 12-volt lamps in each. Color filters are used to balance the spectral distribution to that of the camera sensor for color transparencies and prints.

Production
Udo and Reimar Lenz delivered the three cameras in late January 1998, and the construction of the stands, platens, and light sources, and assembly of the computers, was completed in late February. The system, with two cameras, was installed on schedule in early March in the Prints and Photographs Division. However, the start up phase, scheduled to take two weeks, took about two and a half months. This delayed the beginning of production until late May.

The camera stand, platens, and light source worked pretty much as expected. The masking mechanisms in the platens were inadequate and had to be redesigned. The light sources' DC power supplies were insufficient, and had to be replaced. There were numerous problems with the computer system, but most were resolved during setup. In the rush to get started, JJT failed to align the lens board with respect to the sensor and to properly check the first images to assure uniform sharpness. The client pointed out this oversight, much to JJT's embarrassment. Every image is aligned on the platen so that at least one edge is level with the world, and then rotated when necessary to produce a properly oriented image. Although an image may occasionally get scanned with the orientation reversed and not detected before delivery, this has not created a significant problem.

JTT's Image Capture Station

By utilizing two cameras operating on an 11.5-hour shift, JJT has been able to achieve the onsite production requirements. A third camera, operating in Austin, scans negatives from the Mead-Bateson collection, which is on 5 foot rolls of 35mm nitrate film. Each of the three cameras has a dual Pentium II workstation, supported by 180 Gb NT servers. One data entry station, two inspection stations, one 100 CD-R record station, and one label printer on-site, and five correction and evaluation stations in Austin support the two on-site capture stations. All image operations are synchronized by the central database system.

The following graph presents the number of images captured each day with the two cameras at LC. It illustrates the time it took to get into full production, and indicates that JJT has exceeded the contract specifications to capture 525 images a day since the middle of August. It also indicates that daily production varies widely. The project is on schedule for completion of the first year's work by May 15, 1999. JJT will have scanned approximately 130,000 b/w negatives, 4,000 slides, 5,000 prints, and 300 color transparencies and slides by that time. Work scheduled for the second year is only about one-half that of the first.

Production statistics from May 1998 through March 1999.

Image Quality
Resolution: The initial images, which were scanned at 5,000-pixel resolution along the long dimension, did not meet the Modulation Transfer Function (MTF) values listed in the vendor evaluation section of the contract. As it turned out, these values were impossible to meet in a practical environment at this input resolution. JJT was able to meet the specification by scanning the images at 7,500 pixels, and scaling them to obtain 5,000-pixel images. Although the Library determined that the images scanned at 5,000 pixels were acceptable, project staff felt the improvement in image sharpness obtained by over-sampling outweighed the additional scanning and assembly time required for the larger images. In the process, JJT learned that the use of sharpening, even at low levels, detracts from, rather than improves, the quality of the high resolution image. By scaling the images to 5,000 pixels before writing to file, scanning at 7,500 pixels did not overly impact the writing, processing, or storage times. Hence, all the b/w negatives have been scanned at 7,500 pixels.

Compression: The contract stated that the compression of b/w reference images (640 pixels in the long dimension) should yield an average compression ratio of 10:1. This level of compression resulted in artifacts in the b/w images that were unacceptable to the Library. The compression ratio was reduced to an average of about 4:1 in order to rid all the reference images of observable artifacts.

Moiré: JJT found that well over 70 percent of the photographic images from the HABS/HAER collection exhibited moiré patterns when reduced to 1,000 or 640 pixels along the long dimension. The moiré pattern did not appear in the un-scaled image, but resulted from patterns within the original negative. Project staff went through several iterations before settling on the best combination of blurring at high resolution and sharpening at low resolution to eliminate these patterns with a minimum reduction in image sharpness. Although JJT used an extensive review process, one percent of the low-resolution images delivered to the Library exhibited moiré patterns. In response, the Library approved JJT's recommendation to process all images as if a moiré pattern would exist in the lower resolution images.

Original images produced using the "halftone" printing process represented another source of moiré patterns. Again, the un-scaled digital images did not exhibit moiré patterns because the scanning resolution was sufficiently high to resolve the individual dots. Project staff processed the reference images to eliminate these patterns with a minimum reduction in image sharpness.

Tonal Distribution: Adjustment of the tonal distribution began with the specifications outlined in the contract. These defined tolerance values for white and black areas of a grayscale target, and the rather vague requirement that "care should be taken that no clipping (i.e., loss of details) in either the highlights or the shadows occurs." It is probably fair to say that none of the staff at JJT or the Library knew just what to expect from the high-resolution images. The amount of information captured in them was staggering. The question became one of how best to preserve and display this information.

The initial goal of the project was to ensure that all the important details could be observed on the monitor, including those in the highlights and shadows. JJT assumed this was necessary to confirm that the digital image had captured all the detail present in the original graphic. The on-screen presentation, however, resulted in a rather "flat image." JJT subsequently discovered that while the data may not be present in the displayed image, its presence could be confirmed by referring to the Look up Table (LUT) generated during the correction phase. Adequate tonal distribution was ensured if the LUT satisfied two conditions: 1) the LUT represented the full range of the raw scanned image, and 2) it and its slope were always positive. One can observe the detail of an image generated under these conditions by simply expanding the tonal value as the image is being displayed. The two images below demonstrate the effect of adjusting the tonal distribution.

Project staff moved quickly to an understanding that the most pleasing image possible could be produced within the constraints stated above for the LUT. Hence, correction, evaluation, checkout, and review became very subjective. Each image is viewed by three different JJT staff members before being delivered to the Library. The first persons crops the image, adjusts orientation, alignments, and the tonal distribution. A second person reviews the "corrected" image and makes additional adjustments she feels are needed. The correction and evaluation are performed on a low-resolution (1Kb) version of the raw image. The correction parameters are saved, and used in processing the high-resolution image. Once the image is processed and the derivatives are produced, a third person inspects the image, and either approves it or sends it back to evaluation or capture. The Library reviews the delivered image, and accepts or rejects it. With few exceptions, the Library has not rejected any images for tonal distribution-related issues except for those that were deemed "too bright" in the highlight area. About five percent of the images were rejected in the beginning for being too bright. This figure has been gradually reduced to less than one percent.

Conclusion
The specifications for image quality stated in the RFP were intended as guidelines, and became just that as the work progressed. JJT and the Library had common goals to produce the best-quality digital images possible, with the Library being the final judge of that quality. RLG DigiNews readers can also judge the quality. Many of the images scanned in this project are on one of the following sites:

• http://memory.loc.gov/ammem/ncrhtml/crhome.html
• http://memory.loc.gov/ammem/bbhtml/bbhome.html
• http://memory.loc.gov/ammem/hhhtml/hhhome.html

JJT learned the hard way the need for close communication between its inspector and the Library's reviewer. JJT and LC personnel spent time reviewing images together during the start-up phase, and came to an agreement as to what was an acceptable image. Unfortunately, Library reject rates were higher than anticipated in the initial production phase. The rate was not reduced to an acceptable level (less than one percent) until the periodic joint review of images still in process was reestablished.

This has been a challenging, exciting, and successful project for JJT. The experience proved that very high-quality digital images from b/w negatives can be produced in a production environment. This project also underscored the fact that even with clearly defined requirements and a knowledgeable staff, imaging projects remain quite challenging.

Sue MacTavish, Lockheed Martin
sue.h.mactavish@lmco.c om

Over the last several years, the Department of Defense (DoD) and the National Archives and Records Administration (NARA) have sponsored a series of studies and conferences on the topic of digital imaging and the selection of the most appropriate digital imaging standard for long-term preservation of federal documents. The most recent one, the DoD-NARA Scanned Images Standards Conference, held March 31- April 1, 1999 at NARA's College Park, Maryland facility, was attended by over 90 individuals eager for an opportunity to learn and exchange information on the current status of imaging in DoD and other federal agencies. The program featured an overview of imaging standards, including the types and extent of their use, and the status of selected imaging projects and standards associated with imaging.

Day One
The first day was devoted to addressing the rationale for the conference and presentations by representatives from various agencies describing current implementations and issues faced. Welcoming comments were made by Dr. Ken Thibodeau, Director, Electronic Records Programs at National Archives, and Burt Newlin, from the DoD, OSD C3I, the conference co-sponsors. Then, Sue MacTavish of Lockheed Martin, the conference facilitator and the project manager for the DoD Imaging Standards Policy Support Task Order, kicked off the conference with a presentation on "Why are we here?" She offered the following reasons:

• The use of computers is changing the way government documents are created, accessed and managed. Electronic records, the Internet, and e-mail have become an increasingly large part of the everyday work environment. To improve access, distribution, and interoperability, federal agencies are converting large numbers of documents from paper to electronic digital images. Increased accessibility to the most current data drives the move away from paper records whenever possible. Among these federal agencies, there is increasing interest in receiving NARA guidance identifying acceptable digital image formats for long-term preservation.
• For federal records requiring permanent retention, long- term preservation of digitally imaged records has become problematic. The advantages of digitally-imaged documents are tremendous, however due to the relatively short life cycle of digital image technology (both hardware and software), it is commonly accepted that all formats used today will eventually become obsolete.
• Electronic storage media deteriorate, and the hardware and software systems on which they can be read become obsolete. For an electronic record, long-term preservation requires that as the technology changes, the record be migrated from one format to another, and then verified to ensure no data were loss during this migration. Limiting the number of image formats to monitor for technology change becomes an essential part of a long-term preservation strategy. Identification of appropriate and relatively stable formats is key to success.

MacTavish's introductory remarks were followed by presentations by representatives from various agencies providing updates on "What's going on:"

• Dr. Scott Lackey from the Center for Army Lessons Learned (CALL) reported on how CALL is moving away from acquiring paper and encouraging pure electronic record acquisitions as they provide more utility to end-users. CALL utilizes the DoD 5015.2-STD metadata requirements as the basis for their system. The metadata are linked to the actual electronic or converted record, with all record components managed as one record.
• Bette Mahoney from the Defense Human Resources Activity (DHRA), Joint Requirements and Integration Office (JR&IO), briefed the group on the Defense Personnel Records Imaging System (DPRIS). DPRIS is an OSD initiative to move toward a common operating environment for electronically querying Official Military Personnel File (OMPF) records systems. All of the military services have converted or are converting their personnel records to digital images in the TIFF format, but have utilized dissimilar header structures and TIFF file extensions. Further there is no common indexing and system architecture. DPRIS employs Web technologies to support electronic queries of these disparate OMPF systems, and to speed up search response times. OMPF records are by far the most used, and therefore the most expensive retired DoD records to support. They are essentially archived "forever," and there is a need for identical standards for active, retired, and archived records.
• Updating the group on the status of the government-wide document declassification program was Kirk Lubbes, President of Records Engineering. It is estimated that there are some 40 million pages of information scheduled to be reviewed for declassification. The documents are scanned in TIFF 6, ITU Group 4 compression. Each document is indexed with up to 26 bibliographic fields. The Electronic Document Interchange Standard (EDIS), a voluntary standard for electronic document interchange among Executive Branch agencies, was developed for the declassification project. The standard governs both document metadata and document images that are to be exchanged for purposes of coordinate review, as well as minimum transfer metadata. Digitizing these records facilitates declassification efforts and Freedom of Information Act (FOIA) requests.
• Steve Wehrly of the U.S. Army Publishing Agency provided an overview of electronic information publishing activities in the Army. His presentation covered the myriad of Army publishing media for: administrative publications (including directives); training and doctrine; and technical and equipment. He reviewed the history of Army electronic publishing; its use of the Web; and the Army's "Less-paper" policy. Wehrly concluded with a discussion on linear and non- linear media, electronic publications and interactive electronic publications, and the issues surrounding archiving of these electronic publications.
• Steve Puglia of NARA presented information on the findings and guidelines of NARA's Electronic Access Project. Puglia's data illustrated clearly the fact that it is not the longevity of digital optical media (30 to 200 years), but rather the digital data system's 5 to 10 year systems life that is the critical factor in migrating data to a new technology or system. He concluded that electronic imaging is excellent for access and rapid retrieval, but a poor choice for long-term preservation.

Two more information sessions followed the "What's going on" presentations:

• George Wenchel, of Lockheed Martin, provided a basic overview of the myriad of standards available, and a discussion of the pros and cons of defacto versus dejure standards. Wenchel focused the majority of his comments on TIFF, PDF, and BIIF (ISO standard ISO/IEC IS 12087-5). He emphasized that although there are currently no digital image formats that are acceptable for long-term preservation, the goal is to identify formats that are likely to live longer than others. By selecting such standards, NARA will be able to reduce the frequency of data reformatting required to migrate data through different standards and technology, and thus to minimize the cost of digital preservation.
• Mike Pickard, also of Lockheed Martin, presented data collected in a survey of DoD and selected other federal agency records managers regarding activities and plans in the area of electronic records management. This survey was sent to 35 federal DoD records officers, and a group of selected federal agencies in October 1998. Results were collected from 25 agencies, with a 71% return rate. The goal was to survey the current efforts by the DoD agencies in archiving electronic records, and to determine which electronic formats were currently being used to generate digital images. The top six responses to this latter inquiry were either defacto standards, proprietary file formats, or 'unofficial' forms of approved standards. TIFF was used by 100%, and PDF by 73%, of the respondents that were using digitally-imaged documents.

Day Two
The second day of the conference was devoted to small group discussions and idea generation. The groups were asked to discuss the "drivers" and "roadblocks" to a successful digital imaging program, and who should be doing what, when, and how. The key drivers seem to be access considerations and FOIA. The key roadblocks were costs and lack of management support for appropriate funding in this area. The lack of standards was not seen as a major roadblock. The group felt strongly that a united government voice was needed, with strong NARA leadership and a means of sharing data.

In the afternoon the groups were asked to review and comment on the recommendations found in the preliminary study report, which was disseminated to conference attendees. These recommendations included:

• Image materials in the most stable, uncompressed format available:
• For personnel records: Image using TIFF for archiving, TIFF or PDF formats for access. Convert all current TIFF 4 images to one standardized format.
• For declassified records: Image using TIFF for archiving, TIFF or PDF formats for access. Historically significant records should be converted to paper, microfilm, or ASCII formats.
• For manuals, standards, directive type material: Image using TIFF, ASCII, and ASCII SGML tagged files for archiving; PDF, HTML or SGML formats for dissemination.
• Develop standard header data guidelines for the TIFF image format.
• Work with the Association for Information and Image Management (AIIM) and the American National Standards Institute (ANSI) to standardize TIFF header data.
• Establish criteria for selection of digital images for accessioning in the National Archives.
• Accession digital images that have been imaged in the most stable format available and those that meet the selection criteria.
• Study and evaluate dejure interchange formats for long- term archive acceptance and application in the field.
• Plan for migration of digital images every 3-5 years with a cost equivalent to 50 - 100% of the costs associated with the original imaging project.
• Study and evaluate migration strategies used in digital data archives for application in the maintenance of textual digital images.
• Develop a standard set of metadata for textual digital images using DoD 5015.2-STD, EAD, and Dublin Core as minimum sets.
• Establish guidelines describing metadata that must accompany digital images when submitted for archival accessioning.
• Study and evaluate formats designed for non-textual material (e.g., photography, aerial imagery, x-rays, radar) for compatibility with textual digital image formats in the archive environment.
• Convert documents that require long-term preservation from an application format to an image format for storage.

Most thought that the recommendations should be grouped, but there was no consensus to eliminate any single recommendation nor was there any clear consensus on what recommendations were the most important. One of the small groups proposed an implementation approach:

1. Manage the process (records management, management, and policy)
2. Study, plan, gather information through cost/benefit analysis of entire life-cycle (especially document preparation, searching, and migration).
3. Pick an interim standard during step 2, which will be accepted and supported by DoD and NARA - this will enable the cost-benefit analysis to be conducted.
4. Practice migration and preservation while documents are in active use.

This conference was conducted as part of the Office of the Secretary of Defense/Command, Control, Communications, and Intelligence (OSD/C3I) sponsored Imaging Standard Support Task Order, and was facilitated by Sue MacTavish of the Lockheed Martin Imaging Standard Support Team. The data collected at the conference will be folded into the study's final report, which will be available in late May. The location of this report will be announced in the June issue of RLG DigiNews.

ABCs of Metadata Workshop
May 3, 1999
To be held in Ottawa, Canada, the Consortium for the Computer Interchange of Museum Information (CIMI) and the Canadian Heritage Information Network (CHIN) are sponsoring a one-day workshop on metadata issues facing libraries, archives, museums, digital library projects, and other cultural information organizations.

Tenth DELOS Workshop on Audio-Visual Digital Libraries
Call for Contributions
Submission Deadline: May 15, 1999
Papers are being requested for the Tenth Delos Workshop to be held in Santorini, Greece, June 24-25, 1999. The Delos Workshop will provide a forum for researchers and content providers to discuss and understand the issues that affect audio-video digitization including, metadata (models and management), storage management for continuous data, and multilingual access.

Electronic Imaging And The Visual Arts Beyond The Digital Archive: The Practical Use Of Images
May 19-21, 1999
Sponsored by the Vasari Project, this conference in New York City will focus on digital imaging and bring together those working on digital initiatives from academia, museums, libraries, and research institutions. There will be presentations on digital projects that have been successfully implemented, discussions on the future direction of digital imaging, moving digital images to the printed page, and color management issues.

Using Metadata for Knowledge Management
May 24-26, 1999
To be held in Buffalo, New York, the OCLC Institute on Metadata will be an intensive theoretical and practical exploration of metadata and its applications in knowledge management systems. The workshop will include lectures, structured lab exercises, and group discussions.

Making Sense of Digital Identifiers for Internet and Other Online Applications
June 25, 1999
This one-day pre-conference workshop held prior to the American Library Association annual conference in New Orleans, will focus on understanding and using unique electronic identifiers, such as DOI, URN, and ISWN and the services that are affected by their use.

American Library Association Annual Conference Digital Programs
June 26-28, 1999
To be held in New Orleans, Lousiana, the programs related to digitization include: Selection for Preservation in the Digital Age; Security and the Digital Library: A Look at Authentication and Authorization Issues; Fatal Extraction: Ethical and Legal Issues in Electronic Publishing; and Aggregating the Aggregators: Delivering the Digital Library to the End User.

Archives and Museum Informatics: The Cultural Heritage Informatics Quarterly
Special Issue: Imaging, Visualization And Humanities Research
Submission deadline: June 30, 1999
Papers are requested for a special issue of this journal that will explore the application of imaging technologies in humanities research.

Second Annual International Digitisation Summer School
July 4-9, 1999
To be held in Glasgow, Scotland, participants in the course will examine the advantages of developing digital collections of heritage materials, as well as investigate issues involved in creating and managing access to such collections.

Museums and the Web 99 Papers Now Available
Papers presented at Museums and the Web 99 are now available on the conference Web site. This collection offers an overview of the issues and opportunities facing museums as programs and projects go online.

New Reports Available from the Council on Library and Information Resources
The Council on Library and Information Resources has recently released a number of reports focussing on digital imaging. All are available in HTML or PDF format. They include Digitization for Scholarly Use: The Boswell Papers Project at the Beinecke Rare Book and Manuscript Library, by Nicole Bouché; Avoiding Technological Quicksand: Finding a Viable Technical Foundation for Digital Preservation, by Jeff Rothenberg; Enabling Access in Digital Libraries: A Report on a Workshop on Access Management, edited by Caroline Arms with Judith Klavans and Donald Waters; and Why Digitize? by Abby Smith.

SPARC Announces Grants to Stimulate E-Journal Publication
SPARC (the Scholarly Publishing and Academic Resources Coalition) is offering $500,000 in grants to projects that transform the scholarly communications process. Partnerships and alliances among universities and professional societies are encouraged. The SCI grants will support projects that offer an alternative to the traditional scientific communication process.

Encoding Dublin Core Metadata in HTML First Draft Available
Now available for comment is the first draft of Encoding Dublin Core Metadata in HTML. It was written in response to the need to document current practice while discussion moves forward on data models and XML/RDF encoding. The draft has already been reviewed by the Dublin Core Technical Advisory Committee.

The William Blake Archive Releases New Electronic Editions
Just released by the William Blake Archive are new electronic editions of Blake's masterpieces.
There are now thirty-one copies of sixteen illuminated books in the Archive.

Universal Preservation Format Feedback Requested
Nearing completion is this new proposed standard in the preservation of digital media. Comments on the working draft are requested, particularly on the user and technical requirements.

Colorado Digitization Project Digital Toolbox
This Web site is designed to guide administrators of digital projects to sites and resources on the many issues related to implementing a digital imaging project.

The Northeast Document Conservation Center (NEDCC) Preservation Manual Now Available
NEDCC has made available on its Web site the third edition of Preservation of Library & Archival Materials: A Manual, edited by Sherelyn Ogden. The publication includes some coverage of digital issues.

Treasures from Europe's National Libraries
The Conference of European National Librarians has created a virtual exhibition with descriptions and images of treasures from all over Europe, going back as far as the 8th century. The national libraries have selected some of the most important artifacts in their collections. The exhibition has been designed so that the treasures can be browsed in various ways. It is possible to look at the exhibition materials from a particular country, but they can also be found according to their format, their content, or creator.

European Visual Archive Project
The European Visual Archive (EVA) is a European Strategic project designed to stimulate the use of multimedia information content by digitizing resources in public archives. The aim is to enhance access to European Heritage through digital preservation of archival collections. Photographs from 1900 to the present are being digitized and displayed on the Web. A total of 20,000 photograph images will be available to researchers, and the project will develop policies for description, image quality, metadata, copyright, and pricing.

Early Canadiana Online
Early Canadiana Online has now made available a new interdisciplinary online database containing more than 1200 full-text, searchable titles.

National Endowment for the Humanities Division of Preservation and Access Grants
The National Endowment for the Humanities has just announced their awards for 1999. Funded are a number of projects that focus on digital imaging. These include; a project at the University of California at Berkeley to digitize medieval manuscripts; Cornell University Library, to create workshops on digital imaging for libraries and archives; and the Amigos Bibliographic Council, Inc., to offer digital imaging workshops in the Southwest.

Question:
I am considering saving the master files for our current digital imaging project as JPEG files. Is it possible to create lossless JPEG images if I use the "highest quality" compression option when saving an image as a JPEG file?

Answer:
JPEG (Joint Photographic Experts Group) is an image compression standard designed for compressing color or grayscale images of continuous tone documents, such as photographs and paintings. JPEG is a lossy compression technique, and it relies on several different compression methods. The most commonly used one is baseline JPEG. When an image is compressed and uncompressed, the decompressed image is not quite the same as the original scanned image, as the values of individual pixels are altered during JPEG compression. The degree of loss can be determined by adjusting compression parameters. It is important to note that the exact reduction ratio depends on the composition of the image. For example, at image quality set to "low," it is possible to achieve a compression ratio of 18:1 (original file size: 18 Mb ; compressed file size: 1Mb), whereas at image quality set to "high," this ratio may be reduced to 4:1 (original file size: 18 Mb; compressed file size: 4.5Mb). Unlike a lossless compression scheme (original and compressed/uncompressed images are mathematically and visually identical), even at highest possible quality, baseline JPEG continues to be lossy, causing varying degrees of distortion, such as blocking and other artifacts.

In addition to JPEG, there is also an ISO standard that defines a "lossless JPEG." Lossless JPEG is a completely different compression method that is really lossless. However, it does not compress nearly as well as baseline JPEG. It typically can compress color images with an approximately 1:2 ratio. According to JPEG Image Compression FAQ, lossless JPEG has never been popular. In fact, no common applications support it, and it is considered obsolete.

There is also a new near-lossless ISO/JPEG compression standard called JPEG-LS. It is also referred to as LOCO as the standard is based on the LOCO-I algorithm (Low Complexity Lossless Compression for Images) developed at Hewlett-Packard Laboratories. In near-lossless compression, the error for each reconstructed sample is bounded by a pre-defined value. It is a lossy algorithm in which each decompressed image sample differs from the corresponding original image sample by not more than a pre-specified value. There is usually a very small amount of loss associated with this scheme. Although JPEG-LS provides better compression ratios than the original lossless JPEG, it does not provide the same high levels of compression offered by baseline JPEG. The main feature of JPEG-LS is its superior placement in the compression/complexity trade-off curve. In December 1998, JPEG-LS (Draft International Standard ISO-14495-1) went to final ballot for promotion as an international standard. The public draft of the proposed standard is available as a PDF file at http://www.jp eg.org/public/fcd14495p.pdf. Further information (including a software implementation of JPEG-LS) is available at HP Labs LOCO-I/JPEG-LS Home Page. This issue's Highlighted Web Sites section also provides a useful link to JPEG-related information, including JPEG2000, which is expected to supersede the current JPEG standard.

RLG and Cornell University Library are pleased to announce their collaboration in the production of a monograph on digital imaging, to be published in February 2000. Co- edited by Anne R. Kenney and Oya Y. Rieger of Cornell University, this 200+ page publication will address selection, benchmarking, metadata, enhancing access, quality control, image databases, program management, and digital preservation. In addition to Kenney and Rieger, chapter authors include Paula DeStefano, The Barbara Goldsmith Curator for Preservation, New York University; Peter Hirtle, Co-Director of the Cornell Institute for Digital Collections; Sandy Payette, Digital Library Researcher in the Cornell Computer Science Department; and John Price-Wilkin, Head of the Digital Library Production Services, University of Michigan. A stellar cast of nearly forty experts representing all aspects of digital imaging will contribute sidebars that highlight major issues or synthesize research in a particular area.

As its title suggests, the aim of this work is to translate theory into practice, offering guidance and recommended practices in a host of areas associated with digital imaging programs in cultural institutions. The chapters will be written to support informed decision making, and to assist program managers, librarians, archivists, curators, systems staff, and others with the range of issues that must be considered in implementing and sustaining digital imaging initiatives. The book will also feature research trends that may have particular impact on the field as well as emerging relevant technologies and techniques that apply to digital imaging.

Additional information on the contents and availability of this book will be forthcoming in future issues of RLG DigiNews.

Papers from the Joint RLG/NPO Preservation Conference: Guidelines for Digital Imaging Now Available
Papers from the Joint RLG and NPO Preservation Conference: Guidelines for Digital Imaging have been made available through the conference web site (http: //www.rlg.org/preserv/joint/confpapers.html or http://www.thames.rlg.org/preserv/joint/confpapers.html from Europe). The papers cover guidelines for selection, guidelines for preparation, and guidelines for digital imaging, as well as updates on preservation metadata and digital archiving (digital preservation). Coming additions to the site include reports from the breakout groups, discussion questions and answers, and a list of issues identified by conference participants as the next initiatives to pursue.

Upcoming RLG Forum: Aspects of Digital Preservation and Archiving, Emory University
As libraries, archives, museums, and other cultural institutions increasingly use digital technologies to organize, preserve, and provide access to resources, there is a growing concern over the lack of sufficient standards and concrete guidelines. Also, using these digital technologies requires institutions to develop new approaches and policies for managing and maintaining digital collections. This forum is an opportunity to learn how some institutions address issues such as unstable digital media, format obsolescence, migration and back-up, policies, access, proprietary software, and digital preservation requirements. Hosted by Emory University in Atlanta, GA, it will be held May 20-21, 1999 in the Robert W. Woodruff Library.

For further information, see the forum agenda or contact Sue Marsh Sue_Marsh@notes.rlg.org; 650/691-2376. 

Feature Article
JJT, Inc: http://www.jjt.com
Request for Proposals (RFP97-9): http://memory.loc.gov/ammem/ftpfiles.html

Technical Feature
BIIF : http://164.214.2.51/ntb/baseline/docs/biif/index.html
DoD 5015.2-STD: http://jitc.fhu.disa.mil/recmgt/rma-ps/index.html
Electronic Document Interchange Standard: http://dprc.seas.gwu.edu/dprc5/declassification_products/edis .htm

Highlighted Web Sites
JPEG Information Links: http://www.jpeg.org/public/jpeglinks.htm

Calendar of Events
ABCs of Metadata Workshop: http://www.cimi.org/cimi_institute
American Library Association Annual Conference Digital Programs: http://www.ala.org/events/ac99/index.html
Ar chives and Museum Informatics: The Cultural Heritage Informatics Quarterly: http://www.archimuse.com/publishing/armu.guide.html
Electronic Imaging And The Visual Arts Beyond The Digital Archive: The Practical Use Of Images: http://www.vasari.co.uk/eva/newyork.html
Making Sense of Digital Identifiers for Internet and Other Online Applications: http://www.lita.org/ac99/precon99.htm
Second Annual International Digitisation Summer School: http://www.hatii.arts.gla.ac.uk/DigiSS99/
T enth DELOS Workshop on Audio-Visual Digital Libraries: http://www.iei.pi.cnr.it/DELOS/WORKSHOP/workshop.htm
Using Metadata for Knowledge Management: http://www.oclc.org/institute/metadata2a.htm

Announcements
Col orado Digitization Project Digital Toolbox: http://coloradodigital.coalliance.org/toolbox.html
Early Canadiana Online: http://www.canadiana.org
Encoding Dublin Core Metadata in HTML first draft available: http://www.ietf.org/internet- drafts/draft-kunze-dchtml-00.txt
European Visual Archive Project: http:// www.eva-eu.org/
Museums and the Web 99 Papers Now Available: http://www.archimuse.com/mw99/
National Endowment for the Humanities Division of Preservation and Access Grants: http://www.neh.gov/html/rec_awds.html
New Reports Available from the Council on Library and Information Resources: http://www.clir.org/whatsnew.html
The Northeast Document Conservation Center (NEDCC) Preservation Manual Now Available: http://www.nedcc.org:/manhome.htm
SPARC Announces Grants to Stimulate E-Journal Publication: http:// www.arl.org/sparc
Treasures from Europe's National Libraries: http://www.bl.uk/gabriel/treasures/entree.html
Universal Preservation Format Feedback Requested: http://info.wgbh.org/upf/index.html
The William Blake Archive Releases New Electronic Editions: http://www.iath.virginia.edu/blake/

FAQs
Draft International Standard ISO-14495-1: http://www.jpeg.org/public/fcd14495p.pdf
HP Labs LOCO- I/JPEG-LS Home Page: http://www.hpl.hp.com/loco/index.htm
JPEG Image Compression FAQ: http://www.faqs.org/faqs/jpeg-faq/

RLG News
Paper s from the Joint RLG and NPO Preservation Conference: Guidelines for Digital Imaging: http://www.rlg.org/preserv/joint/confpapers.html
RLG Forum: Aspects of Digital Preservation and Archiving, Emory University: http://www.rlg.org/events.990520.agenda.html

RLG DigiNews (ISSN 1093-5371) is a newsletter conceived by the members of the Research Libraries Group's PRESERV community. Funded in part by the Council on Library and Information Resources (CLIR), it is available internationally via the RLG PRESERV Web site (http://www.rlg.org/preserv/). It will be published six times in 1999. Materials contained in RLG DigiNews are subject to copyright and other proprietary rights. Permission is hereby given for the material in RLG DigiNews to be used for research purposes or private study. RLG asks that you observe the following conditions: Please cite the individual author and RLG DigiNews (please cite URL of the article) when using the material; please contact Jennifer Hartzell (jlh@notes.rlg.org), RLG Corporate Communications, when citing RLG DigiNews.

Any use other than for research or private study of these materials requires prior written authorization from RLG, Inc. and/or the author of the article.

RLG DigiNews is produced for the Research Libraries Group, Inc. (RLG) by the staff of the Department of Preservation and Conservation, Cornell University Library. Co-Editors, Anne R. Kenney and Oya Y. Rieger; Production Editor, Barbara Berger; Associate Editor, Robin Dale (RLG); Technical Support, Allen Quirk.

All links in this issue were confirmed accurate as of April 13, 1999.

Please send your comments and questions to preservation@cornell.edu.

Trademarks, Copyright, & Permissions