Ghosts In The Machine Video - Laboral
Gijón, Spain - LABoral Centro de Arte y Creación Industrial (2008-09).
Video clip (.mov / 00:03:15).
Ghosts In The Machine Video
Mexico City (2008). Video clip (.mov / 00:03:40).
Electronic Voice Phenomenon recordings are made by recording controlled static. Occasionally, one hears sounds that are like human speech. For some the voices are subjective interpretations, for others, the voices are a possible means of communication with the dead.

Ghosts in the Machine references the ideas in EVP to examine ways in which we transform worlds, and bodies in worlds, through pareidolia, apophenia and the gestalt effect. The work uses the strategies of EVP, voice and pattern recognition, and face tracking to generate voices, and images from apparently closed, silent and empty spaces and systems.

This work uses the ideas inherent in Electronic Voice Phenomenon (EVP) to examine ways in which we construct the world through pareidolia, (a psychological phenomenon involving a vague and random stimulus - often an image or sound - being perceived as significant), apophenia (the seeing of connections where there are none) and the gestalt effect (the recognition of pattern and form).

EVP is the recording of errant noises or voices that have no explainable or physical source of origin. These recordings are made when the recorder is alone, or under controlled circumstances. Most often white or pink noise is used as a medium that is acted upon by other electromagnetic forces. This electromagnetic medium produces forms that are, occasionally, like human speech. For some the voices are simply subjective interpretations - that we tend to hear voices in random patterns of sound, in the way we recognize forms in random visual patterns. For others, the voices are genuinely mysterious, opening up the possibility of communication with the dead.

Interest in spirit communication through electronic recording dates back to at least the 1940s and has its roots in the turn of the century Spiritualism movement (1840s-1920s). Originally labeled "Raudive Voices", after parapsychologist Konstantin Raudive, recordings thought to be spirits were later renamed "electronic voice phenomena".

In Ghosts in the Machine (2008) two projectors project large images onto the walls of a room. One projection shows video static overlaid with text and the outlines of bounding boxes, the other shows black and white images of what appear to be blurry and indistinct images of human faces. Ambient noise fills the space. Just at the threshold of recognition can be heard what appear to be human speech in different languages.

A CCD camera is turned on but enclosed in a light tight box. Its input is adjusted with maximum gain and brightness to reveal the video noise inherent in the system. This noise forms the optical equivalent of audio noise and is used in a similar way to provide a medium that can be modified by external forces to produce images and sounds. The video noise is mapped to audio by sampling pixels in a Quicktime matrix and using the values to manipulate a stream of white noise. Voice recognition software parses the modulated noise and translates any sufficiently voice-like sounds into its nearest vocal equivalent, which is sent to the screen as text and rendered into audio by speech synthesizers into Italian, German, English or Spanish.

Face tracking algorithms using a cascade of Haar classifiers scan each video frame and look for any combination of pixels that form the basic characteristics of a human face. These are areas that can be loosely characterized as eyes, nose and mouth with a sufficient degree of symmetry. When the software finds such a combination of pixels and symmetry, the software draws a bounding box defining the area and zooms the area to full screen, its contrast and brightness is adjusted, blurred and desaturated to clarify the found images.

The images produced are only occasionally reminiscent of human faces. More often than not the images produced are recognized as indeterminate organic forms with volume and space, but fail to resolve themselves into anything recognizable. But occasionally, images are produced that are strikingly like a face although in actuality containing only the barest possibility of being so.

In this installation the computer does the hard work of analyzing a complex visual field, but the real task of meaning making is left to the observer. The algorithms find faces in the field that barely meet the requirements of a facial arrangement. The structuring of these images really consisting only of blobs and indeterminate grain, as faces is left to the observer.

Seeing, representation and the interpretation of external phenomena has never been a matter of objectivity. Seeing is a complex activity and the perception of visual forms, aesthetic experience and cognitive interpretation are more at home with the aleatory, the misperceived and phenomena of indeterminacy than with the notion of the world as a fixed reality. It is these that drive the installation Ghosts in the Machine.

Ghosts in the Machine is a generative, closed system. Random noise from a CCD camera is analyzed for patterns. An algorithm looks for patterns that match the basic geometry and physiognomy of the human face. What it actually finds are pixels on a screen forming blobs and patches of colour that have no actual relation to a real world face. They have no indexical relation to an object. They are not images of people, but another kind of image loaded with meaning, which arises accidentally, but irresistibly, from the hybrid interaction between machine and body. To all intents and purposes when these patches of pixels look like faces, they are images of faces. That such obscure images resolve themselves into faces without conscious effort, and that remain even when attending closely to them, suggests that it is paradoxically their lack of objective meaning that generates their form. It is the very ambiguity and intedeterminacy of the images that allows the brain to reconfigure them as indexical.

This work is one of several that examines systems of meaning making that rely on pattern recognition, and the problematized relationship between meaning and the meaningful. The development of meaning in the Projectís work is dependent on an increasingly, yet seemingly infinite, complex recursive and recombinant loop between meaning made and meaning found. In this loop the external and internal worlds are blurrily indistinct, each acting upon the other in the construction of a new self/space forever suspended at the point of becoming. The effort to restablilize the self in this world where everything is in play, is questioned and negotiable, is unavoidably revelatory and reproblematizes current and preceding models of authenticity and resistance.

Using the coincidences of science, para-science, culture, technology and art, the Project examines the fundamentally contingent nature of meaning. The work of the Project has grown out of an interest in bodies in motion, worlds in flux, and in the endlessly recombinant texts and forms of our worlds, and out of an interest in the seen and unseen, the half perceived and misperceived things at the limits of our perception and in the reanimation of the lost bodies and past events that constitute this invisible world. The world is not entirely what it appears to be and that the surface of the visible world needs only to be lightly scratched to reveal the invisible worlds above and below.

[See also: The Sound of Silence]

Ghosts in the Machine (2008)