Materia Prima

The Derniers Souffles installation consists of three industrial boxes that have become reliquaries of electronic circuits. Connected to loudspeakers with compression drivers, each circuit – fashioned from brass, an alloy used in jewelry – amplifies the hum of a component, transistor or diode, emitting a sound akin to white noise. This hum is then modulated to create variations in amplitude and simulate the sound of the tide. The project follows on from a series of workshops with nonverbal autistic youth at special needs schools. During sound naps, the youth were able to achieve states of relaxation previously difficult to attain, through the use of low frequencies and white noise. Compelled by the results, we began to explore the psychoacoustic and suggestive potential of white noise, and ways of producing it. We sought to generate noises with a particular texture, which led us to analog circuits with old germanium (Ge) and silicon (Si) components. We went scavenged, dismantled and salvaged old military and industrial equipment to find, test and compare these once-popular components, which are now obsolete. Our search led us to delve into the history of these components and the conditions under which they were produced, as well as the extraction of resources needed to build and improve them. That, in turn, brought to the fore a number of social and environmental issues related to techno-industrial history, which altered our understanding of the noise. Widely appreciated for its relaxing qualities, white noise paradoxically emerges as the artificial sound of technological excess.

The psychoacoustic properties of white noise

White noise is a signal made up of the sum of the frequencies in the spectrum audible to humans (20 Hz to 20,000 Hz), all emitted at the same level. In other words, all frequencies from low to high are emitted at the same time and at the same volume. Certain frequency bands can be removed from white noise to obtain different variations such as pink noise, brown noise, and so on. While pure white noise is artificial by definition, similar sounds are found in nature, such as the sound of the tide, rain or a waterfall, which are commonly associated with tranquility and well-being.

White noise has long been a source of intrigue. In 1667, Pope Clement IX, an insomniac, realized that he slept better when he heard the rush of fountains. That led sculptor and architect Gian Lorenzo Bernini to invent a white noise machine: a wheel that struck a series of paper globes.¹ While white noise can clearly evoke the sound of a fountain, rain or the sea, it can also generate other, more subjective apparitions, such as voices or melodies. And it's probably because it's made up of all the audible frequencies that our cognitive mechanisms are prone to hear these sound illusions, known as pareidolia. Our brains seek to recognize identifiable forms in abstract sounds, which are interpreted via our imaginations, personal frames of reference, and cultural constructs.²

The relationship between our psyche and white noise has been studied since the 1950s in neuroscience, psychoacoustics, the social sciences, and psychology. Studies have focused on its effects on sleep; schizophrenia; the sucking reflex and the perception of pain in infants; anxiety and its sedative effects; depression; attention disorders; hyperactivity; and its ability to help workers cope with working conditions.³ While the exact mechanisms at play when listening to white noise remain unknown, its positive effects are thought to be linked to two phenomena: its capacity to reduce the perception of surrounding noise, and its properties of stochastic resonance (a phenomenon in which the transmission of a signal is boosted by a noise). All of this is subject to the white noise remaining below a certain volume, beyond which these effects are reversed.

In today’s world of visual and sonic saturation, white noise is becoming a refuge, a balm to calm our nerves, and even a chart-topper that is breaking listening records on streaming platforms, dethroning Beyoncé on Spotify.⁴

The diagram above is from the article “Build a pink noise generator: Cut out noise pollution and keep your cool,” published in the 1970s amateur electronics magazine Popular Electronics.⁵ A decade later, Forrest M. Mims, a science writer who penned many books on electronics, particularly musical, described simple circuits for generating noise in the same magazine. Mims’s work was our gateway into the world of DIY audio electronics.⁶ In this article, he points out the paradoxical nature of interest in white noise, which was initially thought of solely as the random, unwanted part of a signal that interfered with the information transmitted.

“Noise is equally unpopular among radio astronomers, biomedical engineers, radio communications users, and others who work with low-level electrical and electromagnetic signals. So much engineering effort is devoted to the suppression of noise (it can never be entirely eliminated) that it might come as something of a surprise that there are many useful applications for noise. These include acoustics measurements, instrument calibration, antenna tuning, signal jamming, data encryption, electronic music and even applied psychology!”⁷

Nowadays, white noise is found throughout the music industry and music production. It is integral to both the calibration of audio equipment and to composition itself. It can be used untouched, filtered to reduce its frequency range, or modified using effects. It can also be used to add texture, modulate other sounds, serve as a source of random values, or synthesize the sound of cymbals and snare drums in drum machines. In audiovisual and radio production, it for mixing different sound samples. We're constantly listening to white noise without even realizing it.

For Derniers Souffles, we wanted to work on the textures of noise, which led us to tinker with old diodes and transistors that each emit their own specific (and sometimes irregular) hum. These components once figured prominently in technological feats, but are now considered too unstable and sensitive to their environment, especially humidity and heat. Musical creation is one of the last areas where they are still used, as their instability can be used to obtain particular harmonics. When we discovered how the sounds produced by the components resulted from their mineral composition and the industrial purification and manufacturing techniques of their era, we decided to delve further into their history.

The conditions for producing noise

What are diodes and transistors? While both can be made of germanium or silicon, their roles are different. Diodes are electronic components that allow current to flow in a single direction. Transistors have two functions: they can amplify a signal (in a stereo amp, for instance) or serve as a logic gate that switches the flow of current on or off. That was crucial to the development of the binary system in our computers, which only know two values: 1 (the electric current flows) and 0 (the electric current doesn't flow).⁸

Germanium, a metalloid with semiconducting properties, was used in military radar starting in the 1940s. A series of experiments led Bell Laboratories to create the first transistor in 1947. The company thought it had struck gold, but the invention was strategic for the US amidst the Cold War, a time when political tensions were driving an urgent push for innovation. The Federal Government required Bell to share their patent for next to nothing.⁹ The commercial uses initially envisioned (hearing aids, for example) were shelved in favor of military applications such as the radios used in aircraft and supplied to soldiers.¹⁰

In 1954, Texas Instruments developed the first commercially available silicon transistor, which was more precise and less sensitive to external disturbances.¹¹ Silicon transistors replaced the bulky, fragile vacuum tubes used in the first electronic circuits, paving the way for the first miniaturization of electronics while also offering a spectacular increase in performance. The innovation opened the door to a multitude of practical applications, from radio to television and computers. Today, nearly all our electronic devices contain massive amounts of transistors. In 1965, the most powerful circuit had 64 transistors. In 1993, a Pentium chip contained over 3 million, and today’s processors have tens or even hundreds of billions.¹²

In 1965, Gordon E. Moore, an engineer at Fairchild Semiconductor (who went on to co-found Intel three years later), published an article in Electronics magazine. He explained that, since 1959, the complexity of systems using semiconductors had been doubling every year at a stable cost, and projected that this rate of growth would remain unchanged for years to come.¹³ The theory would prove true: between 1971 and 2001, the density of transistors in a system and in microprocessors doubled every year. This exponential increase in the number of transistors per part was achieved through finer etching of circuits on silicon wafers, leading to smaller, cheaper electronic machines that are also faster and more powerful.

Moore's Law is not a scientific law as such, but rather an economic prediction.¹⁴ Since 2007, the physical limits of miniaturization – components now measure a few nanometers – have rendered Moore's projections obsolete. Work is underway to find new ways of combining and layering the ores and metals in components, but each step requires new, more costly processes.¹⁵ Rock’s Law, a newer economic prediction named after early Intel and Apple investor Arthur Rock, states that the cost of a semiconductor plant doubles every four years, as the photolithography manufacturing process used for the past 40 years is nearing its physical limits.

Today, to manufacture a transistor, you need silicon, which makes up 90% of the earth's crust in the form of silicate.¹⁶ Silicon is not “rare” as such, but as with many strategic minerals, it has to be extracted and purified. The process involves high temperatures and, consequently, a lot of coal, electricity and/or gas. It also requires massive amounts of pure water – millions of liters for each plant, along with chemicals such as arsenic, antimony, phosphorus, hydrogen peroxide, nitric acid and sulfuric acid.¹⁷ The chemicals, which have long been discharged into wastewater, are only now starting to be governed by environmental regulations mandating treatment and reuse. Some, but not all, countries now require toxic products to be filtered and acids neutralized.

This has led to an exodus of the resource-hungry, polluting semiconductor industry to Southeast Asia and especially to Taiwan. This massive relocation was spurred by lower labor costs and less stringent environmental standards, and not because the region possesses specific resources that cannot be found elsewhere.¹⁸ Troubled by the shortage of microchips during Covid, the European Union and the United States are currently attempting to safeguard their supply by bringing this strategic industry back within their borders. Environmental considerations have taken a backseat as the big chip manufacturers are now overtaking traditional heavy polluters like the auto industry in terms of their carbon footprint and hazardous waste generated.

We saw parallels between this state of affairs and a phenomenon called avalanche breakdown,¹⁹ which occurs in humming components. Inside diodes and transistors are two plates separated by what is known as a depletion zone. Normally isolated, electrons travel in a direct line between these two plates without passing through this zone. However, if a large amount of current flows through the component, the electrons will be carried along by the oscillations of the silicon or germanium ore and projected into this depletion zone. Their movements then accelerate, until they collide with atoms.

"They then free other electrons: the number of free electrons increases rapidly, because the new free electrons knock other ones, in a phenomenon comparable to an avalanche.”²⁰

This avalanche creates a chaotic movement of electrons that produces a multitude of frequencies: a hum. When amplified, the hum becomes the almost-white noise we know so well. This chain reaction and the chaos it creates seem comparable to the effects of technological development and its deleterious consequences for the environment. We wonder, then, if this “final hum” is not also the last gasp of an industrial society swept along by the wave of its unbridled desire for innovation, and if the sea we think we hear in the white noise is not the sea mentioned in the IPCC report, which estimates that by 2050, around one billion people could be living in low-lying areas threatened by rising sea levels and episodes of flooding during storms.²¹

That is the tension we wanted to address with this installation. The soundscape created by the buzz and hum of obsolete electronic components is an invitation to relax, but it can also be interpreted as a warning of the threats that the expanding technosphere poses to our planet. Derniers Souffles can be read as an allegory of our current situation, suggesting that this soothing white noise may be the last we hear before we are swallowed up.

Over the course of the project, through presentations of intermediary stages, open electronics workshops and exchanges with artists and musicians, we realized that a large number of people shared our desire for emancipation through experimentation. They were interested in an artistic approach based on a reappropriation of electronics and sound, but also in the perceptive dimension of white noise. We felt it was important to find ways to share our Derniers Souffles research so that it could play a part in other creations, so we published a library of samples on the Freesound website under a Creative Commons license. As we work, we collect sounds that we discover from components with interesting textures.²²

These hums of components became the raw materials for three concerts when the project was shown in June 2024. As part of a group dynamic, we took part in a residency at the GMEM to set about creating an instrument with three musician-artists with very different practices: Meryll Ampe, a musician on the experimental scene, Thibaut Quillet, a composer for film, and Simon Cacheux, a creator of sound pieces for museums. Designed to be played by several people, combining control panels and manipulable objects, the instrument will serve as a means of exploring musical gesture and electronic sounds, as well as a source of new collaborations.

Translated by Ethan Footlik

Bio

Sonia Saroya and Edouard Sufrin explore the sensory impact of technological environments and their effects on cognitive mechanisms. Drawing inspiration from the history of science and technology, as well as the properties of minerals and industrial remnants, their installations and sound sculptures engage our reflexes and conditioning, creating paradoxical experiences. By connecting traces of the past with the present, their work seeks to better understand societal evolution while offering moments of escapism. Pursuing emancipation through the conscious use of technology, Sonia and Edouard also share their expertise through educational projects and teaching initiatives.