The following is a guest post by Steve Puglia, Digital Conversion Services Manager for the Office of Strategic Initiatives.
The Federal Agencies Digitization Guidelines Initiative, formed in 2007, includes representatives from various Federal agencies dealing with the digitization of cultural heritage collections. This group meets regularly to discuss technical topics of common interest, with the goal of providing practical guidance on digitization. As is usually the case, there is no lack of topics for discussion. When digitization is done well, we believe the resulting digital collections should meet current and future needs of researchers, as well as enhancing long-term digital preservation and sustainability (see Digitization is Different than Digital Preservation: Help Prevent Digital Orphans! )
The FADGI Still Image Working Group met recently on July 12, 2011, and the following bullet points provide an overview of the topics discussed , as well as the general interests and activities of the group.
- The Still Image Working Group has done a lot of work focused on measuring the performance of digital capture systems, including the development of the Digital Image Conformance Evaluation (DICE) targets and related software. New activities in this area include developing targets for evaluating automatic document feed (ADF) and oversized scanners, and updates of the analysis software. Currently, the DICE device and object targets have a rigid support and therefore can not be used with ADF scanners. So, there is a need for a new target on a thin, flexible support. Also, when evaluating scanners capable of digitizing oversized originals, currently several DICE targets need to be used and analyzed individually. There is an obvious benefit of having different sized targets matched to the size capability of different scanning systems, all of which can be analyzed in an automated fashion.
- In cooperation with CIE Committee TC8-09 Archival Color Imaging, the Still Image Working Group and an international complement of digitization labs are engaged in a study of the color accuracy of current digital imaging techniques. The goal of this study is to identify and recommend practical approaches for improving color accuracy. Organizations will still need to decide how accurate the color imaging needs to be for specific digitization projects. The first round of tests by institutions in the United States is nearing completion, and a second round by European institutions should begin soon. Some initial study results were presented at the IS&T Archiving Conference in May 2011. This work complements research done on color imaging within art museums, including two studies conducted by faculty and staff at the Rochester Institute of Technology and Munsell Color Science Laboratory – Current Practices in Fine Art Reproduction (final report) and Direct Digital Image Capture of Cultural Heritage (final report).
- The File Formats Sub-Group (see Sub-Groups ) is considering the relative advantages and disadvantages of different file format choices. Many institutions still have an expressed preference for uncompressed TIF format for master files, while many other organizations engaged in mass-digitization have opted to use lossy-compressed formats like JPEG 2000 to make the volume of digital data manageable and large-scale projects practical, affordable, and sustainable. What is the “right” answer? The usual unsatisfactory answer is “it depends.” We are discussing five common file formats – TIFF, JPEG 2000, JPEG, PNG, and PDF – and hope to provide guidance on each format’s suitability and pros/cons for different use cases. The goal here is to help people make informed decisions about what file format(s) to use for their digitization projects.
- In support of the file formats discussion and as a follow-up to the JPEG 2000 Summit held at Library of Congress in May 2011 (see A Fine View at the Summit of JP2 ), staff at the Library of Congress are working with a consultant to identify and compare methods for evaluating the effects of lossy compression on image files. This work is in progress, and includes identifying both objective and subject assessment methods people have been using, investigating potential new approaches, and some practical testing of various methods. We plan to report on this work within a few months.
- Finally, the current FADGI web page and the Technical Guidelines for Still Image Digitization of Cultural Heritage Materials will undergo some updating in the near future, to make the information more understandable to less technical audiences, and more usable by everyone. Within the Technical Guidelines, we want to do a better job of expressing the technical concepts and the limits for the various imaging performance metrics in easily understood terms. Also, we plan to make the relationships between the Content Categories & Digitization Objectives and recommendations for information capture and image quality levels more explicit. The Technical Guidelines offer broad recommendations for information capture (such as spatial resolution and bit depth) and ranges for scanner performance and related image quality (expressed as a rating from 1 to 4 “stars”), but currently the recommendations are not associated with particular use cases described in the Content Categories & Digitization Objectives. We believe describing clear relationships between these, based on real world digitization projects, will be a big help for everyone planning digitization projects.
Digitization approaches within the library, archives, and museum community have matured a lot over the last decade. The FADGI Still Image Working Group is playing an active role in these efforts, and will continue to work on improving the technical guidance.
(Update: due to technical difficulties, this blog entry, originally posted on Friday, August 26th, was re-posted on August 29th.)