Julia Kim and Don Mennerich
The Electronic Media Review, Volume Four: 2015-2016
Jeremy Blake (1971-2007) was an American digital artist of time-based paintings or digital animations. While he exhibited Digital C prints and moving images, his working process combined 8mm film, vector graphics, and hand-painted imagery to create his distinctive aesthetic: color-drenched, atmospheric, and even hallucinatory. His work was included in three Whitney Biennials (2000, 2002, and 2004). Blake also achieved rare commercial success in mainstream film and video games with his work on P.T. Anderson’s Punch Drunk Love (2002) and Rockstar Games’s Grand Theft Auto (1997-2013) video game series. New York University Libraries acquired the Jeremy Blake Papers in 2009. His archive includes hundreds of pieces of optical media, Jaz disks, digital linear tape, and multiple hard drives, as well as copies of files from hard drives. The majority of the archival content is comprised of working files with a smaller subset of final artistic works. Over the course of a nine-month span, the co-authors, New York University Libraries Digital Archivist and National Digital Stewardship Resident, tested and implemented workflows for the Jeremy Blake Papers. Working from an archival perspective in collaboration with technologists, curators, and individuals familiar with Blake’s artistic process, this paper documents the Blake case study. It presents some of the challenges, as well as the opportunities, for preserving and creating access to a born-digital collection beset with problems inherent to born-digital work: huge size, problematic formats, layers of interdependency, as well as ethical and intellectual property issues.
Blake’s papers were acquired in two different ways. In 2009, one batch of files was copied from Blake’s MacBook Pro laptop computer, an external hard drive, and a USB flash drive, but the hardware itself was not retained and write-blocking was not employed. Copying the files without a write blocker introduced changes to the source media (Forensics Wiki 2015). This copied hardware was left at the donor’s residence, but New York University Libraries (NYU) acquired Blake’s undergraduate oil paintings and four boxes of almost 400 pieces of media. Much of the born-digital media was in the form of optical disks, with a range of compact disk digital audio (CDDA), DVD, and CDs. The collection also included three external hard drives, as seen in fig. 1.
Starting in fall 2014, the Digital Archivist, National Digital Stewardship Resident (NDSR), and a graduate student conducting an independent study worked to image or copy this media as an additional batch (Kim 2015).
Methodology was multi-fold. We drew upon emerging archival processing practices in complex born-digital media, while also borrowing from digital forensics and computer science. This was supplemented with principles and practices from conservation ethics, literature, and documentation practices.
Safeguarding the media through practices such as re-housing the physical material, photographing, and creating bit-exact copies for retention were taken from the beginning of the project. However, the Blake Papers also necessitated analyzing the content with an understanding of methodologies in place for artwork. This concern for authenticity drove strategies that included emulation, migration, and inviting scholars, art historians, and technologists for collaboration and interpretation. The specific nature of the unfinished work that comprises the bulk of the Jeremy Blake Papers also influenced workflow strategies.
In the initial stabilization of the collection, digital forensic methodologies and tools were employed. The Forensic Toolkit (FTK) is an example of such a tool. FTK is a powerful software program that allowed the entire collection corpus to be analyzed dynamically by filtering for dates, media-types, file types, and much more (fig. 2). FTK also served as a preview environment, supporting the preview of hundreds of file types including moving image and image files.2
Throughout the course of processing, a comparison of checksums confirmed a significant proportion of duplicative files. That is, there were multiple, bit-exact copies of the same files throughout the collection. Blake’s working method included frequently backing up his files.1 This, in itself, is neither unique nor interesting, but through comparing hash values of duplicative files, overwritten metadata lost in the copying of files on-site at the donor’s home were recovered (see fig.3).
The duplicative files make it easier to understand where, how, and how often Blake backed up his files. Future researchers will not only be able to look at the files themselves, but will now have the tools to gain a holistic understanding of how Blake worked with his files and his hardware to create his finished works.
THE PHOTOSHOP FILE
While Blake’s animations were created in AfterEffects, the collection was primarily composed of still image file types, specifically Adobe Photoshop files, or PSD files. To create his work, Blake painstakingly collaged and overlaid many different sources into Photoshop. He drew images digitally on a Wacom tablet, scanned found materials, generated complex vector graphics, and inserted his own photography. These were further manipulated and processed in Photoshop. To create the finalized moving images, he directed close collaborators to work from his layered Photoshop files. The Blake Papers include explicit directions for collaborators, with desired timecodes and visual effects in word processing files and through the use of textual overlays in Adobe. The heart of the collection and the key to understanding his artistic process lay in the PSD files.
Photoshop, despite its prevalence in art-making and collections, presented new and surprisingly little understood issues (Rimkus et al. 2014). According to Adobe Photoshop’s available documentation, the Photoshop file itself is proprietary and incompletely understood (Adobe Photoshop 2013). Therefore, despite its widespread adoption, authorities such as the Library of Congress do not recommend it as a preferred format for long-term preservation (Library of Congress 2013). One resulting complication was that identifying the version of the PSD files was problematic. Standard file identification software failed to consistently differentiate software versions. This was further complicated by Blake’s own unorthodox practice of using the layer functionality in Photoshop to create what are, in effect, key frames for reference in animating a sequence. As seen in fig. 4, each layer is highly distinct and necessary for understanding Blake’s working processes (Mennerich and Kim 2015).
While some file-based preservation strategies advocate for normalization, normalizing to Tagged Image File Format (TIFF) 6.0 presents major compromises. According to TIFF documentation, TIFF 6.0 files do not support layering. Normalization to TIFF 6.0 specification would render a file with over 50 distinct layers—that can be turned on and off in different ways—into a singular flat image.
The preservation strategy had to be multifold and go beyond saving the files as is. The three-prong strategy involved migrating the files to the most up-to-date version of Photoshop using Adobe Photoshop Bridge, while also normalizing to TIFF 6.0, despite the lack of support for multi-layered TIFFs in the TIFF standard (Mennerich and Kim 2015).
While there is surprisingly robust backward-forward compatibility with Photoshop files across Photoshop software versions, there is no guarantee that this will always be the case. We were also concerned with researcher access to something akin to the original creation environment, which would have included extra slow processing time, access to content aggregated on the optical media level, and the ability to explore multiple versions of Photoshop within an operating system. As part of our methodology, comparisons with contemporaneous machines and programs (fig. 5) were made to monitor differences and loss encountered with the tested emulations (fig. 6).
This was balanced with a pragmatic exploitation of the native range of variability inherent in Blake’s born-digital work. We strategically recognized that this digital media art, while not quite mass culture, involved a range of creation and access platforms. While authenticity and analysis of significant properties was a concern, this was countered with embracing a more sustainable, less onerous approach in keeping with our larger goal for creating workflows useful to other collections. Free from the exhibition responsibilities of a museum or gallery, the authors were able to experiment with how researchers access works-in-progress. Displaying finished works would have necessitated working with an understanding of exhibition histories, and exacting requirements for display monitors and color settings. Finalized moving image formats would have presented a much simpler case study with established best practices and target formats.
After several weeks of trial and error with proprietary software programs and open-source emulators, we found the PSD files compatible with an Ubuntu 10.04 operating system (released in 2012) on an older laptop with a version of SheepShaver, a popular PowerPC Mac OS emulator supporting versions 7 – 9.0.4. Major incompatibilities, dysfunction, and instability were observed in trials with non-Linux operating systems and other contemporaneous operating systems. Further study remains to be done.
This is a pioneering case-study at New York University and its complexity has challenged many of our preconceptions on how to best handle these types of materials. Just as the use of write blocking hardware and software has only in recent years become de rigueur, the practices we employed were, at times, guided by an attempt to find not a necessarily a best practice, but a working practice to enable a degree of access and preservation to the artwork and process. This project also set into motion further work with piloting access types with a small group of researchers. This case study is neither complete nor exhaustive; further work and collaboration across disciplinary and institutional lines is called for, especially in the areas of open-source emulators, the Adobe Photoshop file, and understanding researcher access to this type of material.
The authors would like to acknowledge the generosity of and thank the George Stout Fund for sponsoring the inclusion of this presentation at the 2015 AIC Annual Meeting. The work would not have been possible without the funding and support of the National Digital Stewardship Residency. Additionally, special thanks to Lisa Darms and Amy Brost.
1. Blake expressed unconcern for the problems with preserving his work: “The digital file theoretically can stay new if you keep copying it, although I don’t know if that is true. But what do I care? It is the museum’s job to catch up and I am not worried about the museums. It is not what I can do for the museums, it is more what can the museums do for me. But a lot of artists might have forgotten that at this point, too. Why not make something that could survive a nuclear war at that point or the flood in New Orleans? It is just ridiculous. You do what you want to do (Teine 2012, 130).”
2. Peter Chan’s short video tutorial on the adaptation of FTK into archival workflows explains this much more clearly (Chan 2011).
Adobe Photoshop. 2013. File Formats Specification. www.adobe.com/devnet-apps/photoshop/fileformatashtml/ (accessed 9/15/2015).
Aware Systems. 2008. TIFF, Tag Image File Format, FAQ. www.awaresystems.be/imaging/tiff/faq.html (accessed 9/15/2015).
Borghoff, U., P. Rödig, L. Schmitz, and J. Scheffczyk. 2006. Emulation: Current Research and Development. In Long-term Preservation of Digital Documents: Principles and Practices, Berlin: Springer-Verlag. 212.
Chan, P. 2011. Processing Born Digital Materials Using AccessData FTK. Stanford University Libraries. www.youtube.com/watch?v=hDAhbR8dyp8 (accessed 9/15/2015).
Forensics Wiki, The. 2015. Write Blockers. www.forensicswiki.org/wiki/Write_Blockers (accessed 9/15/2015).
Kim, J. 2015. Jeremy Blake’s Time-Based Painting. NDSR-NY Blog. http://ndsr.nycdigital.org/jeremy_blakes/ (accessed 9/15/2015).
Library of Congress. 2015. Recommended Formats Statement. www.loc.gov/preservation/resources/rfs/stillimg.html (accessed 9/15/2015).
Mennerich, D. and J. Kim. 2015. Preserving Jeremy Blake’s Archives. Archiving 2015(19): 49-50.
Rimkus, K., T. Padilla, T. Popp, and G. Martin. 2014. Digital Preservation File Format Policies of ARL Member Libraries: An Analysis. D-Lib Magazine. www.dlib.org/dlib/march14/rimkus/03rimkus.html (accessed 9/15/2015).
Teine, S. 2012. Jeremy Blake’s Time-Based Paintings: Sodium Fox: Fragmented Crypto Self-Portrait. Saarbrucken: LAP LAMBERT Academic Publishing GmbH & Co.
B.A., Columbia University
M.A. in Moving Imaging Archiving and Preservation, New York University
Folklore Specialist and Digital Assets Manager, American Folklife Center
The Library of Congress.
Washington, DC 20002
MS in Information Systems, Pace University
MLS, Simmons College
Digital Archivist, New York University.
New York, NY 10012