Digitizing Images at the Digital Labs
My work at the Brooklyn Museum involved the digitization of film, transparencies, and glass slides and negatives from across different collections. I focused on the physical scanning and editing of slides and negatives using an Epson Epson Expression 10000XL Flatbed Scanner and Adobe Photoshop; metadata aggregation of these images using Excel and The Museum System database; creating bibliographic records and batch processing images and metadata using the Museum's digital asset management system Luna; uploading the images to the Brooklyn Museum's browsable online digital collection.
Phase One: Digitizing Transparencies from the Library Collection
For the first half of the internship, I worked on the digitization of transparencies and film negatives form the Library's collection of magazine and print materials.
The transparencies were first reviewed by Deirdre Lawrence, the Principal Librarian of the Brooklyn Museum, and flagged for digitization. Before scanning begins, I have to make sure the transparencies have not already been scanned and processed by previous interns. This is also the time to read through the training manual and become familiar with naming conventions, scanning standards, and information used to create metadata.
A typical work flow looked like this:
The first step includes scanning images using an Epson Expression 10000XL Flatbed Scanner, following specific workflows created by the Digital Labs staff. Adhering to these standards ensures that all digitized images meet the miminum quality standards for the Museum's digital collections.
Example 2: Scanned black and white glass slide, before color inversion
Once scanned, various photo edits are performed: images are cropped, rotated, inverted and color-corrected to ensure the best possible presentation in the online digital gallery. This step requires a sharp eye and the ability to stay as true to the original image as possible.
Example 3: Inverted image, before color correction
Images are saved as uncompressed TIFF files, using a standardized naming convention that varied depending on the origin of the collection (for example, Library collections vs. Objects collection). The TIFF files can be very large and are ideal for preservation, but they are compressed into JPEG files in the presentation side of Luna (the public side of the online digital gallery) for various reasons, including copyright concerns and ease of display and thumbnail creation.
Next, we create a spreadsheet that contains the metadata of the images: the filename, information of the original art item and negative slide properties, batch load information, and tracking data for proper internal management of the images and their metadata. Data used in this step comes from a variety of places: the digital file itself, for technical and structura metadata, The Museum System database for administrative and descriptive metadata. This is a surprisingly time consuming process, as it is at this point where all the research and quality assurance (triple-checking of information) occurs. It's only after aggregating the most complete and accurate metadata possible that we can begin the process of creating metadata and mapping files to be used in batch processing.
Example 4: Metadata spreadsheet
To upload digital objects and metadata to the database using Luna DAMS, you need three things: the image files, the metadata file, and the mapping file. The mapping file connects the images to the metadata, so it is incredibly important to make sure there are no typos or formatting errors, otherwise the images will not connect to the metadata.
Once we are ready to upload the images and metadata to Luna, we follow the workflow batch process in Luna, and within a couple of hours all the images uploaded will be displayed in the Brooklyn Museum's Digital Collection website:
Example 5: Screenshot of Brooklyn Museum website, with actual scanned image
Phase Two: Digitizing Glass and Film Negatives of Works of Art in the Museum's Collection
This project followed the same digitization protocol as above, with a few noted differences.
Firstly, the majority of the slides were 8x10 glass negatives, which are incredibly fragile and therefore take longer to process, as only one slide can be digitized at a time. In comparison, the slide transparencies from Phase One were each 3x4 and could be scanned in batches of 6 at time, resulting in a faster processing time.
Secondly, because the glass slides belong to the Objects collection, they required use of a different schema for metadata creation. This presented a new learning challenge, as I had to learn a new standard that was markedly different from the standard used by the Library's transparencies collection. This is a great example of how different collections from one institution can require completely different processes for digitization, which adds time and complexity to digitization projects. The oftentimes subjective nature of different collections and their digitization needs is something all librarians and archivists should take into consideration when planning proposals and project outlines for digitization projects in their respective institutions.
As the Fall 2010 semester winds down, I am still working on this phase of the project, and project to finish by January of 2011. This will conclude over 120 hours of time spent learning about the digitization process at the Brooklyn Museum's Digital Labs.