The digitization of the collections housed in the Charles H. Templeton Music Museum represents a significant emphasis and priority for the Mississippi State University Libraries. The well-preserved collections represent one man’s distinct yet diverse interests in music and business. They span a timeframe from the Civil War to the Vietnam era, and cover all genres of popular music during that time. The collections document the evolution of America’s society, economy, international policy, and especially, its music.

The Sheet Music Collection was first appraised and arranged/categorized by David A. Jasen, musicologist and internationally known author, composer, performer, teacher and collector. Mr. Jasen assigned each piece to one of 7 categories: blues, Irving Berlin songs, foxtrots, movie tunes, minstrel songs, popular, rags, show tunes, specialty, and war songs. He then alphabetized the sheet music by title. However, the collection was not truly cataloged. Therefore, the goals of this project are to catalog, preserve and make the collection accessible. For that reason, every printed score will be inventoried in the form of adding a barcode number on a sheet of acid-free paper along with the Title, Composer, Publisher, Date, and physical location of the piece. This is then inserted into Mylar sleeves with the original piece. The inventory number, description of the physical piece and electronic piece are then recorded in the metadata for that title. This process will insure that every item in the physical collection will be cataloged, regardless of whether or not it was selected for digitization.

As the goals of the project are to catalog, digitally archive and make accessible via the web all titles in the sheet music collection, selection criteria is minimal. While Mr. Templeton did not collect multiple copies the exact same pieces, the collection does contain multiple print versions of the same musical scores. These pieces differ in the cover art, advertisements, and often the performer featured on the cover. Though each version may have its own merit as a unique piece because of the social and historical relevance of the different features, they are considered ‘duplicates’ for our purposes.

Preservation is a primary goal of the project, and certainly we would prefer to digitally archive every item in the collection, but due to the limited resources available we have chosen to scan only unique titles for this phase. For now, however, when multiple versions of a score are encountered, selection is based primarily on the physical condition of the pieces. Typically, one version is in better condition to be handled and scanned: one version is faded more, or has tears or creases or other undesirable attributes. In these obvious cases the best-preserved piece is always chosen.

In other cases where the differences are in the covers or the advertisements, the criteria are more subjective. We generally chose pieces with cover illustrations rather than text-only covers, unless the physical condition of the item makes it less desirable. The cover art provides a glimpse into popular culture of the times, and has value to researchers and other patrons who may have little or no interest in the sheet music itself.

The methodology described here was adopted in January of 2007, when the production phase of the U.S. Small Business Administration funding began.

For each piece of sheet music scanned, the following files are being created:

• Archival copy scanned at 600 dpi and saved as a TIFF files stored on a separate share on the digital media archive. Plans for FY 2007/2008 include installing a separate SAN to store the archive quality images for long term preservation.
• One folder per score containing all archival TIFFs.
• Print-quality PDF files at 200 dpi.
• Thumbnail image (JPEG, 72 dpi, 120 pixels in height) of the cover.
  

Scanning for this project is done with Epson 1500 Gt-15000 flatbed scanners at 600dpi (dots per inch) resolution, which meets or exceeds the best practices recommended by most digital projects. The covers are scanned in 24-bit color, while the scores and advertisement pages are scanned in 8-bit grayscale.

We have chosen to adopt the ‘scan-once’ philosophy, which dictates that all items should be scanned and archived at the highest practical resolution, creating archival quality digital surrogates for the printed scores. This process will help prevent any situation where the original pieces would need to be handled.

The scanned images are then saved with no alterations as uncompressed TIFF files. These are considered the ‘digital archival’ versions. Each of these institutions differ slightly in their standards for capture and processing, as well as their format(s) for delivery. The constant evolution of software, hardware, and storage media make standards for this type of project very difficult to nail down. After reviewing the advantages and disadvantages of the other collections, we decided to err on the side of caution for the archival images, and scan everything at 600dpi. The high scanning resolution and the fact that the files are stored using no compression make for large files, and consequently requires more storage space. Our decision to error on the side of caution and scan at 600dpi was based on the fact that we do not want to allow the storage issue to prevent us from setting high standards for archival images and because electronic storage devices are increasing in capacity and becoming more affordable.

The “delivery,” or “access” files present more of a challenge. Most sheet music digital collections wrestle with the issue of accessibility versus printability; is it more important that the files load quickly in a web browser, or print legibly? For many sites the result is a compromise which does not adequately meet either goal. We chose to focus on the print quality of the sheet music, as bandwidth and file transfer speeds are constantly improving. To this end we create a PDF document comprised of the cover and all pages of the original print item, including advertisements. The access files are consequently large and can take a few moments to open, but CDM plans to release a software upgrade in fall of 2008. Much of the focus of the upgrade will be on increased support of PDF files. The new version should extract pages from PDFs to load individually, and therefore more quickly, while retaining the settings of the PDF so that print-quality is not compromised. If the upgrade does not improve the access time to within a few seconds, we will have tested other options for providing separate display and printer-friendly files. This in-house solution would be more time-consuming and therefore costly, so we plan to wait to see how well the CDM version 4.3 PDF enhancements work before proceeding. In addition to the sample pieces we’ve provided, we are also providing links to access prototypes of the scores as we expect them to display after the upgrade. The scanned images are compressed to 200 dpi, sized to 8 ½ by 11 inches and compiled into the PDF booklet using Photoshop image processing software. The covers are not otherwise altered, but the interior pages, including the sheet music itself, are enhanced for better printing. The background field is converted to pure white, and then any smears or smudges that might obscure the print are removed using the brush tool. Any stamps, handwritten notes or other markings are left intact, to preserve the authenticity of the original collection.

The data for this project is stored on a Windows 2003 ii6 server in a secure location in the Library where it remains unaltered and password protected. This allows our scanning technicians to copy the archive copies to it, but not remove or change them. The images are digitized from desktop workstations and uploaded directly to the server. This server will use a Storage Area Network with a redundant copy of all data. Keeping the data on a live disk in this manner ensures that the data is accessible and does not rely on CDs or DVDs which have an untested shelf life. Tape backups to LTO3 media are performed at regular intervals to add a second layer of storage and copies are stored off-site for additional protection in the event of a disaster.

• Digital Library Federation- U.S. National Archives and Records Administration (NARA) Technical Guidelines for Digitizing Archival Materials for Electronic Access: Creation of Production Master Files: http://www.diglib.org/pubs/dlf103/


• Library of Congress Technical Standards for Digital Conversion of Text and Graphic Materials: http://memory.loc.gov/ammem/about/techStandards.pdf


• Cornell Digital Imaging Tutorial: http://www.library.cornell.edu/preservation/tutorial/contents.html


• Recommended Best Practices for Digital Imaging of Musical Scores: http://www.dlib.indiana.edu/~jenlrile/oclc/oss.pdf


• OCLC- Recommendations for Digitizing for RLG Cultural Materials: http://www.rlg.org/en/page.php?Page_ID=220


• Digital Imaging Guidelines for Hudson River Valley Heritage: http://www.hrvh.org/about/digitalimaging.htm


• University of Southern Mississippi Guidelines for Digitization: http://www.lib.usm.edu/~spcol/crda/guidelines/index.html#capture


• A Framework of Guidance for Building Good Digital Collections: http://www.niso.org/framework/Framework2.html


• Mississippi Digital Libraries Guidelines for Image Capture: http://www.msdiglib.net/guidelines/capture.htm