As Patrick wrote in the panel proposal:
"Increasingly digital media are stored and compressed in ways such that no data is lost. Such files are called lossless, an evocative term that points to the fantasies of total preservation and perfect reconstruction that reach back to the earliest days of 'new media.' [...] Indeed one might argue that losslessness is the principal fantasy of historical production, that is, the notion that we might reconstruct a perfect representation of the past through those objects that remain."
Henry Lowood will be serving as our respondent. We are also honored to be sponsored by the Media, Science & Technology Scholarly Interest Group.
Panel Number: J14 | Thursday, 5:00-6:45pm
Chair: Patrick LeMieux • University of California, Davis
Jacob Gaboury | Stony Brook University
A Voice Uncoiled: Noise, Loss, and the Origins of Digital Sound
Patrick LeMieux | University of California, Davis
Losing Time: Microtemporal Histories of Speedrunning
Laine Nooney | Georgia Institute of Technology
What the Engine Can’t Reveal: The Other Lives of Sierra On-Line’s AGI Development Software
According to the 2014 Nielsen Music Report last year in the United States over 9.2 million vinyl records were sold, a growth of 52% over the previous year. This trend is part of a growing interest in and preference for analog media over their digital counterparts. We might interpret this alternately as an investment in those sensible and material traces inscribed into physical media by analog signals, or as nostalgia for what we perceive as a more authentic form of mediation. Yet in the early-1970s when digital audio was first being developed and commercialized it was precisely these traces that engineers and computer scientists sought to silence or otherwise erase. Digital audio allowed for the fantasy of total abstraction, of a recording divorced from the medium or context in which it was produced. While analog media often leave traces of their material specificity in the signal they capture or record, digitization allows for the transformation of sound into information, uncoiled into its constitutive parts. In the 1970s this was cutting edge technology, yet its origins lie in an act of preservation and restoration, an attempt to recapture the past through digitization.
This paper traces the origins of digital audio in the 1960s and 70s with particular focus on the work of Thomas G. Stockham and his Salt Lake City-based company Soundstream Inc., the first digital audio recording studio in the United States. Its primary case study is the Caruso Project, an early attempt by Stockham to digitally restore a series of gramophone recordings made by opera singer Enrico Caruso for the Edison Company from 1902 to 1920. Developing a process known as “blind deconvolution,” Stockham was able to digitally clean the recordings of the noise of the acoustic horn that funneled the sound of Caruso’s voice to the recording mechanism. This work had wide reaching implications, as it allowed for the restoration of any signal without knowledge of the original conditions of its production. The Caruso Project would subsequently lead to the commercial development of digital audio and sound recording, real time digital editing, and the production and standardization of the compact disc format. Through an investigation of this early history this presentation looks at what is lost through the dual processes of restoration and digitization, and asks what place those messy artifacts that characterize analog media have in the history of technology.
Losing Time: Microtemporal Histories of Speedrunning (Patrick LeMieux)
Sitting alone in the bedroom of his childhood home in Quincy, Massachusetts and livestreaming Super Mario Bros. (1985) to a modest audience on Twitch TV in the middle of the night on July 1, 2013 Andrew Gardikis shouted, “Oh my gosh! I don’t even know! Oh my gosh. It’s gonna be like, it’s gonna be like 4:58.03 or something!” After all this time, could it actually be a 4:57? Gardikis does not know because, in that moment, he cannot know. The temporal precision of speedrunning far outpaces the manual dexterity necessary to accurately hit a split or stop a watch. Speedrunning is the practice of playing video games as quickly as possible. Super Mario Bros. - the 1985 platformer hit from Nintendo - is one of the earliest and perhaps most competitive games to speedrun. Between 2007 and 2014 the world record for Nintendo’s classic game dipped from 5:00.60 to 4:58.09, and with these incremental losses the figure 4:57 became the almost mythical horizon of possibility for human play and Gardikis’ ultimate goal--what he calls his “gaming masterpiece.” After each attempt to beat the world record Gardikis must decode and time his DVR recording frame by frame in order to uncover his time in retrospect. At this very moment when time is most important, it is simultaneously lost. Likewise, when the frames are finally counted, the act of play is lost.
The problem of loss--of the inability of record times to adequately represent the complex material practices of speedrunners--is highlighted by Gardikis’ plight and by speedrunning more generally. Related to both Henry Lowood’s (2006, 26) “high-performance play” (i.e., “play as performance, modification of content, and community-based tools and content development”) and James Newman’s (2008, 124) “superplay” (i.e., “the use of the knowledge and techniques uncovered and laid out in Game Guides, the exploitation of the structures, [non-] linearity and limitations of videogames as designs as well as the harnessing of glitches in game code”), speedrunning encourages players to adopt temporal constraints in order to convert single-player games into massively-multiplayer metagames in which time becomes the central rule governing play. As a result, a popular history of speedrunning is almost always represented as a series of numbers—an asymptotically plunging line graph reducing years of community research, races, and records to a single set of digits. This presentation details the microtemporal history of speedrunning Super Mario Bros. and theorizes the desire for and difficulty of reducing kinetic play to abstract numeral--what could be called a “lossy ideology.”
What the Engine Can’t Reveal: The Other Lives of Sierra On-Line’s AGI Development Software (Laine Nooney)
This paper is an archival and theoretical analysis of the Adventure Game Interpreter [AGI], a proprietary 1980s in-house game engine created by home entertainment software producer Sierra On-Line. The AGI engine was developed from 1982-1984 and was implemented as the company's standard game development platform throughout the 1980s. Accounts of this transition document the engine's significance as largely graphical and intrinsically progressive: it afforded the move from static image adventure games to an exploratory, interactive “2.5-D” game space. However, as this paper argues, adoption of the AGI engine was not simply an elegant programming maneuver; it was a critical infrastructure bound up in three specific realms beyond the immersive experience of the games themselves. First, the AGI engine fundamentally altered puzzle design; graphical space became a constitutive feature of an adventure game, and thus design practice itself at Sierra changed. Second, the engine became the backbone of financial strategy at Sierra On-Line, proprietary software allowing the company to limit development costs while increasing output. And third, it inadvertently unshackled programming from the necessity of mastery within the company's programming department, permitting “levels” of expertise to emerge and formalize.
In the context of this research, AGI is a case study challenging the reigning presumption within media archaeology and materialist media studies that code is the defining condition of a game object’s ontological status. Embedded in this analysis of the AGI engine is a methodological rejection of the theoretical fantasy of code and platform as inherently lossless, prior to and more material than the environment of its arrival. Underpinned by work in feminist science studies and feminist media archaeology, this presentation offers a case that critiques the concept of the supposedly self-evident programmatic layer of an artifact like a game engine, instead emphasizing the co-relational character of the engine. As I argue, code is often privileged precisely because it is merely a materiality problem; for all the unique preservation issues it creates, it is simpler to identify than the messier, entangled phenomena of people, affects and practices bound up within the engine’s own existence. Thus, this paper contributes both original archival research on game history and a methodological provocation concerning the status and boundary of a media object.