This photograph shows a three-dimensional representation of sound using paper. The spectrum (frequencies from low to high) is represented by an arrangement of single strips of paper, with lower frequencies in the foreground. The changes in the spectrum over time are visible as variations in the profile of the paper strips, if read from left to right. Such paper models were used at the Technische Universität Berlin in the 1960s to represent the sounds of speech (phonetics) and music (acoustics). In the background of the photo is the score of the recording analyzed, the first eight measures of Beethoven’s Symphony No. 8. At the beginning of the score, all the instruments of the orchestra are playing, including bass instruments and timpani; accordingly, low frequencies are represented in the first part of the waterfall plot. Later, only the woodwind section is playing, accompanied by French horns, resulting in a decrease in low frequencies in the second part of the plot (to the right). The sound recording to be analyzed was filtered to include a very narrow band of the spectrum, and this signal was visualized on a strip of paper using a level recorder. Individual strips were then glued onto cardboard, cut out, and arranged in a row. This process was repeated for each band of the spectrum, resulting in a three-dimensional spectrogram.

The creation of these “waterfall plots” (Frequenzgebirge in German) is described in Musik…, verwandelt: Das Elektronische Studio der TU Berlin 1953–1995. A picture on page 99 of that book also shows a large waterfall plot made in a different way: the individual strips of paper represent spectra measured consecutively, not the variation of single bands.