IST E-infrastructure shared between Europe and Latin America Volcano seismograms sonification application Domenico Vicinanza, CERN EELA Conference, Santiago, September 04-05, 2006

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Volcano eruption forecasting Currently no definitive method to predict the eruption of a volcano has been discovered or implemented (yet). Scientists monitor –seismic waves –number of earthquakes and the intensity of a specific type of quake (harmonic tremors) in the run up to eruptions. –changes in the shape of the volcano or concentrations of gases emitted from the cone.

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, The hope: music for forecasting By correlating spectra and melodies with precise stages of volcanic activity we hope to discover a sort of “signature tune” of an imminent eruption or earthquake. By identifying musical patterns that warn of an eruption it would be possible to implement civil protection measures, days or even hours before the event

IST E-infrastructure shared between Europe and Latin America Data audification can be considered as the acoustic counterpart of data graphic visualization, a mathematical mapping of information from data sets to sounds. Data audification is currently used in several fields, for different purposes: – science and engineering, – education and training, mainly as data analysis and interpretation tool. Although most data analysis techniques are exclusively visual in nature, data presentation and exploration systems could benefit greatly from the addition of sonification capabilities. Data Audification

IST E-infrastructure shared between Europe and Latin America Sonic representations are particularly useful when dealing with complex, high-dimensional data, or in data monitoring tasks where it is practically impossible to use the visual inspection, or for pattern detection. Research has shown that people are quite more confident in recognizing patterns audibly rather then visually Music theorists and researchers have carried out in centuries of history lots of techniques and methods to detect, study and classify musical phrases Main idea: Music as a language and music analysis as a tool to inspect scientific data Motivations

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Advantages of hearing Hearing is non-directional It is surely easier to recognize a change in a sound with respect to a modification in something which has to be looked at. It is quite impossible to distinguish a blinking light flashing 100 times a second from another one flashing at 200, 1000, or times a second, while anyone can recognize periodic signals from 20 Hz to (almost) Hz

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Sonification on the GRID network First experiments involving sound production with INFN-GRID facilities started during the last months of In September 2003, it was installed CSound, a free and cross-platform acoustic compiler, on a GRID test site, the Catania INFN-GRID computer farm The compiler was tested within the new environment and since its beginning, the test phase produced interesting results: efficient use of the calculus resources, customizable quality of the audio files.

IST E-infrastructure shared between Europe and Latin America After a four-five months test phase, CSound has been distributed to other GRID farms in Europe, Data sonification, sound synthesis and algorithmic composition have been investigated using couple of Csound files and Python (or sometimes Perl) scripts. Python scripts have been used to prepare the right data files during the tests. First Phase: CSound

IST E-infrastructure shared between Europe and Latin America Second test phase: development of a sound production suite based on Java (equipped with the standard audio and math libraries), more flexible and easy to manage. All the results presented in this website have been carried on using this last approach: sample computation, audio rendering, DFT computing were obtained with the Java sonification program on the GRID Executable = "/bin/sh"; StdOutput = "sonification.out"; StdError = "sonification.err"; InputSandbox = {"sonification.sh", "Sonification.java", "etna.dat"}; OutputSandbox = {"Sonification.aiff", "Sound.dat","Spectrum.dat", "sonification.out","sonification.err", "logfile"}; RetryCount = 7; Arguments = "sonification.sh"; Second Phase: Java

IST E-infrastructure shared between Europe and Latin America Sonified data were geophysical data collected by digital seismographs placed on the Etna volcano in Catania (Italy) and on Tungurahua volcano in Ecuador. We carried out two sonifications: – seismogram straight sonification (tranformation into an audible waveform) – seismogram melodisation (tranformation into a melody) Sound form volcanoes Tungurahua (Picture: M. Monzier IRD/IG-EPN) Etna Volcano

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, About seismograms audification In both the cases, structural properties of the seismographic information would be straightly mapped into sound or melody properties In the first case, regularities in the seismograms will be reflected by the existence of spectral lines in the sonified signal In the second case, regularities in the seismograms will be transferred into regularities in the melody (such as a repeated set of data will become a repeated musical phrase)

IST E-infrastructure shared between Europe and Latin America Seismographic data have been recorded onto the surface of volcanoes, by digital seismographs at a sampling frequency of about 100 Hz ( Hz). Seismograms sonification

IST E-infrastructure shared between Europe and Latin America Example of ASCII files processed (sample from Etna data): Starting time: 15/06/ :03: Frequency: Hz Samples: Original data

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, First Sonification: From data to waveform Scaling procedure to properly arrange the samples in the [-1,1] interval, according to their sampling frequency. Users can specify in the sonification program a certain resample frequency (pitch shift). Setting resample factors in the Java code greater than 1 won't preserve the original pitch, allowing a frequency shift making audible regular phenomena happening at very low frequencies. In this way it is possible to observe and study periodical patterns, regular behaviors, long-range correlations, which can happen at different time scales.

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Sample and hold and linear interpolation It have been implemented two interpolation methods to generate the waveform: linear interpolation and sample-and-hold (SH) interpolation. Linear interpolation returns particularly "smoother" sounds. Nevertheless, for some analysis application, it could be useful to prefer the SH audio file, where no external values are added to the original (scaled) data set.

IST E-infrastructure shared between Europe and Latin America Creating the waveform: Linear interp.

IST E-infrastructure shared between Europe and Latin America Creating the waveform: Sample & Hold

IST E-infrastructure shared between Europe and Latin America The waveform coded in the audio file will have exactly the same regularities, also recognizable thanks to the presence of some higher lines in the spectrum. The order of magnitude of the frequency of quasi- regular phenomena is in the range 0-50 Hz, with a spectral envelope centered around Hz. Quasi-regular phenomena 20x resampled Etna seismogram

IST E-infrastructure shared between Europe and Latin America Etna waveform and sonogram

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Tungurahua Waveform

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Spectral lines = Regular patterns Tungurahua Spectrum

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Tungurahua Sonogram Time evolution of the spectrum. Each vertical slice is the spectrum at a certain time Oscillation pattern variations are clearly visible in the pattern

IST E-infrastructure shared between Europe and Latin America Etna sonification within GILDA

IST E-infrastructure shared between Europe and Latin America

IST E-infrastructure shared between Europe and Latin America

IST E-infrastructure shared between Europe and Latin America

IST E-infrastructure shared between Europe and Latin America

IST E-infrastructure shared between Europe and Latin America

IST E-infrastructure shared between Europe and Latin America Listening to sonification.aiff file Etna sonogram

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Data melodization The whole data set interval is mapped on the (equally tempered) piano keyboard The min value of the seismographic data will correspond to the lowest playable note on the piano keyboard The max value to the highest playable note

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Main advantages Already available tools to manage MIDI files and analyze them Tracking the evolution of the musical intervals, the dynamics of their patterns, it is possible to detect, with an high level and in a customizable way, any kind of modifications in the shape of seismogram. Two example of MIDI analysis (free) software: Rubato ( MIDI Toolbox (

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, MIDI Toolbox

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Dynamic evolution of the tonalities (related to couples of adjacent values)

IST E-infrastructure shared between Europe and Latin America Original data Example: Sinusoidal behavior Graphical representation

IST E-infrastructure shared between Europe and Latin America Sinus melodization Music representation (of the same set of data) Data are periodic, so the melody is periodic, with the same period

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Melodization: a pictorial view Pictorially we can say that the melodization works by overlaying seismograms with music notes To create the volcanic score, we take a seismogram and trace the shape on to blank music bars. Then we overlay the contours with musical notes.

IST E-infrastructure shared between Europe and Latin America Seismograms Melodisation

IST E-infrastructure shared between Europe and Latin America Seismograms Melodisation

IST E-infrastructure shared between Europe and Latin America Seismograms Melodisation

IST E-infrastructure shared between Europe and Latin America … have you ever heard a volcano playing a piano ? Seismograms Melodisation

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Melody follows the shape of the oscillation

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, First Etna-Tungurahua duet Players: Mt Etna: Piano Mt Tungurahua: Guitar

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Reference sites: Etna Sonification website: Tungurahua Sonification web repository: MIDI Toolbox manual: nual.pdf nual.pdf

IST E-infrastructure shared between Europe and Latin America Santiago, Chile, EELA Conference, Thanks! …Questions