“What’s wrong with Art?” (1997), by German-Californian artist Stephan von Huene, was subjected to a comprehensive acquisition documentation. To counter the risk of loss of the computer’s specific hardware, a low-budget logic analyzer was used to record and document the sound installation’s digital signal output. This collaborative project was conducted by the author during a fellowship at the ZKM | Center for Art and Media Karlsruhe in collaboration with Daniel Heiss, computer scientist, as well as the museum’s team of electronic technicians
Collecting institutions are faced with several obstacles when aiming to preserve either functionality or future access to time-based media art, particularly when dealing with compiled code, inaccessible electronics, and obsolete hardware. The aspect of time is especially relevant for computer- and software-based art, for which hardware and software define the timing score. The computer-controlled sound sculpture “What’s wrong with Art?” (1997) by Stephan von Huene (1932–2000) is a striking example of compiled code and custom-built, dedicated hardware counteracting preservation or complicating emulation attempts. Additional documentation methods are necessary to preserve the core of its inner logic.
The installation consists of three wooden sculptural constructions with organ pipes colored red, yellow, and blue. Stephan von Huene realized the technical setup. He assembled the computer himself and equipped it with specific plug-in cards (e.g., SMARTLAB, 8-channel relay output adapter) and the MS-DOS 6 operating environment. Batched scripts and compiled code operate an organ blower and electric tone valves for each organ pipe inside the sculptural tower. Voice recordings (WAV audio files) are played out to prelude the organ’s tones. Each score, with a play time of about four and a half minutes, is triggered by the viewer through a radar sensor.
Assessing the risks of loss, future access, and preservation options, it became apparent that the computer—with its individual plug-in cards and compiled code—could not be reactivated, reproduced, or emulated if it failed. For this reason, documentation of the full score of the 8 controlling signal computer outputs was carried out with the aid of a logic analyzer.
Logic analyzers can sample multiple digital signals or channels at the same time, recording every channel’s state (low or high) by bit (0 or 1). Logic analyzers are used in software engineering to test complex circuits, to detect glitches, and to debug digital systems. Logic analyzers bring along a number of key specifications, including channel count, sample rate, trigger modes, and memory depth. In general, two clocking and sampling modes are used in practice: asynchronous (timing) and synchronous (state) acquisition. Asynchronous clocking uses the internal clock of the logic analyzer while the state mode may be synced with another timing source. The gathered data can be visualized and assessed as a waveform time/voltage level graph (e.g., using an open-source processing script). The obtained data can be used for hardware emulation as well—for instance, to replace a system’s functionality/signal by a printed circuit board and a microcontroller.
The 8 digital signals were captured with an Arduino Due microcontroller using 8 of 54 digital input channels and a sampling rate of 1 MHz. For the Arduino setup, the artwork’s voltage level of 13.8 Volts had to be replaced by 3.3 applied volts. The 8 signal streams (all events from high to low, from low to high) were captured through event-based interrupt triggering and were logged in parallel with Δ timestamps in microseconds resolution.
The sample rate plays a key role for timing accuracy when sampling any signal. A fundamental rule for analog signal sampling is defined by the Nyquist-Shannon sampling theorem, which says that a signal can be reconstructed lossless only if the sampling took place at least twice as fast as the highest frequency of the signal. Nevertheless, much higher sampling rates are recommended in practice, especially for digital signals. Therefore, it is important to determine the specific characteristics of the signal, as the present frequency stipulates the required sample rate of the analyzer.
In the present case, the captured signals of the 8-channel relay plug-in card are based on milliseconds due to the shutter speed of the magnetic relays on the plug-in card. However, the signal is sampled with a sample rate of 1 MHz, which leads to the conclusion that the sampling rate is a thousand times higher than the signal under test. Therefore, the logic analyzer Arduino Due solution with its specific sample rate was sufficiently suitable for the present case.
Logic analyzers prove to be very useful tools for the documentation of digital signals in the field of electronic media art preservation. Recording the digital signal output can be a complementary documentation method both as a supplementary visualization and a record of the logical signal stream. Furthermore, the sampled data can be used for active preservation measurements, for example, hardware emulation. Nevertheless, it is essential to consider the appropriateness of a logic analyzer for a specific application. In the present case, a low-budget, open-source logic analyzer was a suitable solution. However, a more advanced logic analyzer might be necessary when dealing with distorted signals or contemporary systems, which reach higher signal speeds.
ZKM | Center for Art and Media Karlsruhe: Daniel Heiss (firstname.lastname@example.org), Christian Nainggolan, Felix Mittelberger, Dirk Heesakker, Marc Schütze, Mirco Frass, Nahid Matin Pour, Katrin Abromeit, Morgane Stricot, Rainer Gabler, Benjamin Miller, and Simone Logos.
Petra Kipphoff von Huene, Jesús Muñoz Morcillo, Martin Obrist, Christoph Noetzli, and Nicholas Popovic.