mechArchival

Wow! Has it been over six months since I last posted? Pretty bad. An update is definitely in order…

With my early November presentation proving to be a resounding success, I was able to spec and order equipment around the first of December.  Nearly everything was delivered during the week between Christmas and New Year’s. Equipment list is thus:

• Canon EOS 5D Mark II camera body
• Canon 24-70mm zoom lens
• Canon 50mm macro lens
• Kaiser Fototechnik R1 copy stand with 5519 baseboard, 4409 column, and RTP arm
• Kaiser Fototechnik 5558 lighting unit
• Canon remote shutter release and 120V camera power supply
• Sekonic L-358 light meter
• Dell Optiplex 380 desktop with two 19” monitors
• Adobe CS5 Photoshop and Photoshop, Camera Raw, and DNG software app
• Small stuff: Gray card; Tiffen/Kodak Q13 color sep. guide; Bulb air-blower; Black velvet cloth; three 30”x60” folding tables; rolling office chair; steel shelving unit; uncoated piece of framing glass; kitchen-type step

The setup is in a long narrow room, with the three tables arranged in a U along the walls at one end of the room. Copy stand/camera sits on the center table, computer on one side and open table for staging material on the other side. I’ve removed the bulbs from the fluorescent ceiling light fixture to avoid conflict with the daylight temp fluorescent bulbs on the copy stand. To compensate I have a small task lamp that I can use when working on the computer. The black velvet covers the copy stand base, creates a neutral background in the image, and produces a high contrast with the edges of most material, which is needed when running the Photoshop auto-crop routine.

Getting things up and running took very little time. Then, starting with settings outlined in documentation published by the Beineke Library, I spent several days shooting different types of materials under differing conditions in order to arrive at a combination of f-stop, shutter speed, and exposure compensation that produces a consistent image quality. I also worked out the post-processing routine with the Adobe apps and gathered information on final file sizes.

Once I was producing what I felt were consistent and accurate images I put together another presentation for management focused this time on gaining approval to put the process into production. I included several back-to-back comparisons showing images produced by our current scanner process with images from the same object produced by the rapid capture process. I also had examples of textual, photographic, and graphic materials of various sizes. Here’s one example of automotive product literature that I produced for the presentation: Catalog from the Electric Vehicles Company. Finally, I included preliminary numbers for master file size and production rates. This presentation well, and the process was approved for production.

I’ve now had a little over six months of experience with rapid capture, with three months of creating production images. Overall, I’m pretty satisfied with how the process works and the resulting output. I do have some lessons learned to share, which I’ll detail in my next post. In the meantime here are a couple of related writings:

– OCLC report, Rapid Capture: Faster Throughput in Digitization of Special Collections, by Ricky Erway. The project site which includes a link to the report in pdf format is http://www.oclc.org/research/activities/capture/default.htm. Many thanks to Ricky for including my comments in the report.

– Article, Digitizing the Collection, in the current edition of Time Traveler magazine, published by The Henry Ford and available online at: http://www.thehenryford.org/timetraveler/ Provides insight as to how our digital material is being used, and features a fantastic picture of me hard at work.

– Brian

On Tuesday I gave a presentation to management on Rapid Capture digitization. I’m recommending we adopt Rapid Capture to make headway on the 180,000 items that have been identified to be completed in the first 24 months of our digitization project.

The Rapid Capture methodology has been adopted or tested by a number of institutions in the last 2-3 years, including the Yale Beinecke Library, the Art Institute of Chicago, The Getty, the San Diego Air & Space Museum, and the University of Minnesota Libraries. In addition OCLC has a study underway to explore where and how Rapid Capture is being utilized.

Information from these institutions is available from a variety of sources. Several Rapid Capture presentations were given at the MCN2009 conference in a session titled Speed the Plow: Rapid Capture Digital Workflow. Minnesota is blogging about one of their projects. The Yale Digital Coffee group has posted a few presentations. And Chris Edwards of the Yale Beinecke has posted details about the RIP workflow used in his Digital Lab.

At its most basic Rapid Capture is a copy stand approach to digitization, employing the new generation of full-frame 35mm digital SLR cameras (e.g. Canon EOS 5D Mark II), batch processing of images, and division of the imaging and metadata streams. Much like Henry Ford created the high volume industrial assembly line through division of tasks, creation of parallel processes and the offering of a simplified but very useable end product, Rapid Capture can potentially generate large numbers of digital images by applying some of the same principals. Low equipment cost is another attraction, with a complete capture station being only a few thousand dollars more than a scanner workstation equipped with an Epson 10000XL for example. The slightly higher cost is balanced by the potential to digitize at similar resolution and at rates 4-6 times than that possible on a flatbed.

At a higher level Rapid Capture involves debate about creation of digital images to primarily facilitate user access and discovery versus creation of images for primarily for preservation of the original object. Fortunately for me, my management is already sold on access, so this turned out to be a non-issue. What may become more of an issue is how we deal with storage of the images we intend to create. If we plan on retaining our physical materials after digitization, do we also then need to retain the uncompressed digital master or just the low res compressed service copies? I don’t expect this particular question to be resolved for some time.

Decisions about the amount and type of metadata created for each image greatly affect the rapidness of Rapid Capture. The Yale Beinecke requires a ratio of 3 catalogers to 1 imaging tech to reach a scan rate of around 400 images per day with the amount of metadata they have chosen to generate. If you operate solely in an archival environment and your images can be tied to the finding aid context perhaps all that is needed is an image number, with researchers developing their own descriptions. In this case the imaging tech may be able to generate all the required metadata directly with no additional cataloging support required. On the other hand if you are operating in a LAM environment there may be the need to generate extensive metadata in order to satisfy needs of a diverse staff, with an accompanying trade-off between staffing level and imaging rate. As with the issues around preservation I expect to be discussing metadata for quite a while.

Based on the response to my presentation we expect now to move forward on acquiring equipment and developing an initial workflow. Stay tuned….

– Brian