»In Search of Good Ancestors / Ahnen in Arbeit« is an experiment in generative radio meant to modulate the attention of the listener, to allow the savoring of slow speech and audible nuance. The stream features an artificial broadcaster, a bespoke voice synthesis system that utilizes current deep learning techniques in voice synthesis and style transfer to be able to shift fluidly through speech and song, words and soundscapes, reading the text as if tracing slowly over the letters, probing them and unsettling them. The concept of the work explores the idea of long-term thinking, and how this idea is wrapped up in the endeavors of prediction, risk analysis, and memory that happen at many levels of human society and consciousnesses.
In the spirit of this theme, the musical composition is governed by a »zoo« of simulation algorithms, essentially, complex systems that cannot be solved »analytically«, through abstract reasoning, and therefore must be run computationally through imaginary time, as simulations. Classic examples of such simulations include climate forecasts, stock market predictions, dynamics of human brains, and the Lotka-Volterra equations, which simulate the life cycles of predators and prey within an ecosystem. The broadcaster performs texts produced by a predictive process. Beginning with a small set of texts, including Jonas Salk’s 1977 lecture »Are we being Good Ancestors?« and Chief Oren Lyon's 1992 address to the United Nations, it attempts to predict what words should come next based on its previous memories, an approach to mathematical future-prediction called »autoregression.«
Commissioned as part of the CTM Radio Lab, led by CTM Festival and Deutschlandfunk Kultur – Radio Art / Klangkunst in collaboration with Goethe-Institut, ORF musikprotokoll im steirischen herbst, Ö1 Kunstradio, and The Wire magazine.