Data-Driven Music and Installations

Sonifying Data Centers

stay tuned! We are working on more projects!

Data centers provide critical infrastructure needed to house servers, data storage drives, and network equipment that power the internet, process, and distribute data, and train, deploy, and deliver AI applications and services. While they provide these services to people throughout the world, they carry drastic and focused social and environmental consequences. These span from causing local power bills to become more expensive as they stress the electrical grid, cutting the water supply to local communities, taking up large amounts of land mass while providing few jobs or resources, such as food or other goods to communities, pollute air the air, and are known to bring emit a constant, audible low hum that disturbs local wildlife and people living close to them.

“all because we can” is a reflection on the rapid investment into and construction of data centers since the AI boom that began in the late 2010s. The piece features a steady counterpoint of data sonification that superimposes the number of data center facilities that have been active, under construction, and announced from 2000, projecting to 2048, alongside their total power capacity from those years.

The city I live in, , and more broadly, are considered hot spots for AI data centers because companies can acquire land, water, and electricity at low costs. The people who profit from these data centers and/or access them to generate text, music, etc, are often far removed from the social and environmental costs that these technologies take. I often wonder if the environmental and ecological impact is worth having the convenience of LLM’s or other generative AI usages that require large models.

“all because we can” for trumpet, interactive visuals, and data sonification (2025-2026)

Alexandria Smith (data driven music) and Kyle Smith (Visuals)

Implemented in third-order ambisonics, the work engages with parameter-mapping sonification strategies and spatial audio techniques to communicate the environmental and societal consequences of unregulated rapid expansion. This composition was authored and data was mapped using Max MSP.  The sound engine was designed in Ableton Live 12 using custom software instruments and  IEM Plug-in Suite was used for spatialization). For the live performance on trumpet (with silent brass mute), I use the Max ICST external to perform spatialize my sound live. This composition can be performed ambisonically, in 8.1 or in 4.1.

The visual layer is staged as a megalithic facade that fills the frame and functions as a brutalist abstraction of resource availability. What reads as solid and immutable at first is gradually depleted as rising demand steadily consumes the sense of available resources. They serve as a guide alongside the improvised performer to communicate the effects and material consequences of resource depletion that this rapid expansion will cause if not mitigated.

The visuals were implemented as a real-time OpenGL Shading Language fragment shader in TouchDesigner. All data was received locally via Open Sound Control (OSC) via the the data sonification max patch. The amount of planned and ongoing construction introduces directional tearing via anisotropic warping. Facility count increases the density of the erosion detail, and power demand biases the cutoff as an additional stress term.