Graphical Sound The Multzvuk group [Vibroexponator](https://120years.net/the-vibroexponator-boris-yankovsky-russia-1932/) "From 1930-34 more than 2000 meters of sound track were produced by the Multzvuk group, including the experimental films ‘Ornamental Animation’, ‘Marusia Otravilas’, ‘Chinese Tune’, ‘Organ Chords’, ‘Untertonikum, Prelude’, ‘Piruet’, ‘Staccato Studies’, ‘Dancing Etude’ and ‘Flute Study’. The Multzvuk archive was kept for many years at Avraamov’s apartment, but destroyed in 1937." “In 1932 Boris Yankovsky has separated from Avraamov’s =ultzvuk Group. He established his own laboratory in Moscow and created essentially new techniques of spectral transformations and synthesis of various sounds, including a singing voice and speech, confirming Avraamov’s foresights on future possibilities of speech and vocal synthesis. The result was an independent control over the pitch and duration of sound on a spectral level, resembling the popular computer music techniques of cross synthesis and the phase vocoder. Yankovsky wrote : “[...] I found the idea of synthesis while I was laboriously working on ‘drawn sound’. And this is the chain of my considerations: The colour of the sound depends on the shape of the sound wave; Graphical representation of the sound wave could be analyzed and represented as the Fourier series of periodic functions (sine waves);” (Zakharine and Meise, 2013, p. 168) “Consequently, the sound wave could be re-synthesized back with the same set of sine waves. Nobody did this before the invention of graphical (drawn) sound just because there were no technical means and methodology for sound reproduction from such graphical representations of sound. As with electrons (the neutrons and protons) the number of which defines the quality of the atom, so do sine waves define the quality of the sound—its timbre. The conclusion: why not initiate a new science – synthetic acoustics?” (Zakharine and Meise, 2013, p. 169) “Right from the start Yankovsky intended to work with a modified animation stand called the Vibroexponator, shooting still images of artificial drawn sound waves by means of a rostrum camera. This meant that the discretization of time scale was predetermined by 24 frames per second, with each successive frame containing one stable sample—a sort of momentary photograph of the constantly changing sound. The audio waveform of each successive frame was calculated according to the spectral content, based on the specific set of spectral templates he had devised. To perform complex mathematical calculations of waveforms as well as other important parameters of sound and automated musical performance such as rhythm and dynamics, there were special ‘employee-computers’ on the staff in the laboratories of Boris Yankovsky and Evgeny Sholpo. These were mathematicians whose specific task was to make calculations (fig. 3).” (Zakharine and Meise, 2013, p. 169) “Yankovsky proposed the method, based on research on structural similarities and distinctions in spectra of sounds of different character to limit as far as “possible the number of calculations needed for synthesis of various complex sounds by means of additive synthesis. In order to achieve this he decided to: analyze the spectrums of various sounds divide all sounds into classes according to common features of timbres, related to spectrums and spectral dynamics analyze these common features (formants, for example) divide the spectrum into groups of overtones responsible for the specific character of the timbre calculate and draw the waveforms, related to these spectral groups build a library of drawn waveforms for further manipulations in the framework of various synthesis tasks. Yankovsky named these final drawn waveforms ‘spectro-standards’ or ‘spectral templates’, semiotic entities that could be combined to produce sound hybrids, based on a type of spectral mutation. According to Yankovsky, for the final stage of sound synthesis there is no need to deal with spectrums but rather to manipulate the predefined waveforms—the readymade templates—selecting them from the library. These templates could be modified by changing their sizes by optical means (or mathematically and graphically) according to calculated desirable amplitudes and frequencies, mixing the results optically or by means of mathematical calculation of the final waveform.” (Zakharine and Meise, 2013, p. 170)