QCD-Audio

Sonification is a new method of data display and is defined as the use of audio to convey information. It is an alternative and complement to visualisation, and proved to be particularly useful in the analysis of highly complex and multi-dimensional data sets. In recent years, it emerged in different fields, pushed by the development of real time audio synthesis software and the increased consciousness of human audition. The growing amounts of data in society and science ask for new approaches in data analysis and display. One of these rapidly prospering fields is computational approaches to problems in physics that have been developed into a powerful paradigm. The rapid development of computers and algorithms has led to new quantitative and qualitative insights, but the typically multi-dimensional data sets are very large, and only a few simple observables are considered. Our interdisciplinary proposal QCD-Audio suggests applying sonification techniques to data of numerical models in physics. Sonification provides a series of powerful approaches for the display and exploratory analysis of complex data structures. Firstly, audification is the most direct approach of sonification, where higher dimensional data structures are mapped into a one-dimensional string of values, which is played as an audio wave. As the ear provides a very sensitive measure concerning dynamic changes, we can listen to highly compressed data and perceive subtle differences in various sonic parameters like timbre and pitch. Secondly, different data dimensions may be used to control attributes of a sound: the combination of pitch, loudness, different aspects of timbre, localisation, (micro)- rhythm, texture qualities etc. results in characteristic auditory streams. These gestalts may easily be recognized and even distinguished, if played in parallel. This parameter mapping allows for displaying high dimensional and dynamically evolving data. Several other techniques of auditory display are applied in science, as model-based sonification, auditory graphs or earcons and auditory icons. Their usefulness for QCD-Audio will also be studied, and new methods of audification and sonification will be developed. During the predecessor research project "SonEnvir - Sonification Environment", relatively simple discrete spin models of the Ising- and Potts type were sonified. In the current project, we plan to continue this line of research. The first new challenge we address here is the step from discrete to continuous spin models (XY model). Such models exhibit topological structures that are harder to capture with classical analytic methods. A next case study will be quantum electrodynamics (QED), as a preparation for the final data we want to sonify: data stemming from lattice QCD - quantum chromodynamics, the theory of the most elementary particles known today (quarks and gluons). This data is given in 4 dimensions. Again, topological structures are known to exist, but they are hidden under violent short ranged quantum fluctuations. It is hoped that sonification will provide a new filtering method for the long ranged topological objects. Within SonEnvir, sonification has been applied to a wide range of problems from neurology to sociology to physics. Many new sonifications methods were developed in recent years, and are shared freely among the members of SonEnvir and the Community of Auditory Display (ICAD). Thus, development in one application of sonification might turn out to be useful also in other disciplines. On the physics side, the Institute of Physics at the University of Graz will provide data and expertise. The Institute for Electronic Music and Acoustics will provide expertise in sound synthesis and rendering and infrastructure for high-end interactive auditory display (e.g., multi- channel sound system including a motion tracking system). The aims of QCD-Audio are, on the one hand, new insights in the data and new perceptualisation methods in the field that might fuel auditory display in many scientific disciplines. On the other hand, besides being an interesting form of display of scientific data, sonification is often used as a basis for electronic compositions and media art. In the line with this branch of new emerging artistic production, an outcome of this project will be a multi-channel interactive sound installation, which allows specialists and amateurs to engage in actual research data of lattice- QCD.

How should elementary particles sound like? Can we hear a difference between different states of lattice quantum chromodynamics (QCD)? These and similar questions were addressed in the project QCD-audio. It examined the use of sonification in the context of computational physics. Sonification is the translation of information to auditory perception, excluding speech itself. Thus it serves as analogue to visualization in the auditory domain, and can be applied to scientific data where visual tools often fail due to huge and multi-dimensional data sets. Human auditory skills of pattern recognition open the floor for data exploration and the formulation of new hypotheses. The data in QCD-audio were provided by simulations of computational physics. This field shows rapid development in algorithms that provide solutions for analytical problems, e.g. in particle physics. One of the studied theories is lattice QCD. The data sets resulting from such simulations are steadily growing with increasing computational power, and thus provide complex test cases for developing sonification tools. For instance, the use of digital waveguides makes it possible to hear size and shape of clusters in so-called Polyakov loop data from lattice QCD, featuring clusters in the percolation state; topological structures in the XY spin model automatically come to the foreground when applying a sophisticated phase modulation technique; granular synthesis facilitates an `acoustical overview` of different phases of the Ising model; and in collaboration with ALICE experiment at CERN, the measurement process of one of its detectors, the time projection chamber, was made audible. These and more examples show how the developed sonification tools are applied. The tools can be transferred to data from other scientific domains. In the project we also argued for metaphoric sonification, where the resulting sound is synthesized in a way to be intuitive for the final users. For instance, the sound propagation in caves is a metaphor for clusters in the QCD example mentioned above, and physicists at CERN were asked how elementary particles should sound like. In order to enhance the communication of sonifications, notation modules have been developed that link the mathematical formulation of data to the one of sound synthesis. As further outcomes of the project, an interactive audio installation using a motion-tracking system was set up for the general public, the `data listening space`. It facilitated the spatial exploration of data from quantum electrodynamics. The workshop `Science by Ear 2` was organized, where an objective comparison between different sonifications was tested, helping to answer the question what a good sonification actually makes out. Listening examples (or demo videos) and the code can be found at www.qcd-audio.at.