Shirohisa Ikeda

Immanuel Kant proposed in 1760 that our knowledge of the outside world depends on our modes of perception. As one’s pinnae are unique, and the filtering they impose on sound directionality is learned by individuals from early childhood natural ear spacing or “head-shadow” of the head and ears, happen naturally as a person listens, generating their own ITDs (interaural time differences) and ILDs (interaural level differences). Based on those individualities this research interest lies in the detectability and intelligibility of the resulting sounds (e.g., Doll, Hanna, and Russotti, 1992; Ebata, Sone, and Nimura, 1968; Plomp, 1976) from the sonification of non-linear non-stationary data (E.g. magnetic fluctuation data from satellites).

Using sonification in our exploration and human perception the basis of expansion for non linear and non stationary space plasma data, is posteriori-defined. A major problem limiting the utility of visual displays is that data sets typically contain much more information than can be effectively displayed using currently available technologies. It is important to consider the limitations imposed by nature on the human eye. For example, even the best computer screens available today are limited to a range of spatial resolution. These limitations affect the useful dynamic range of the display, reducing the amount of data scientists can study at any one time. Scientists currently work around this limitation by filtering the data, so as to display only the information they believe is important to the problem at hand. But since this involves making some guesses about the result they are searching for, many discoveries may be missed.

The interstellar medium is too thin for travelling waves or sound to propagate, it is possible to look at each value measured by the many spacecraft as a sound value. Sound has a number of attributes which can be used for data representation, including pitch, loudness , rhythm, duration and repetition, phase, to name a few. Integrating the latter to current data analysis techniques/visualizations may also us to perform a more detailed analysis of space plasmas. Moreover, we not only propose using sonification as a mean to display information, but to use sonification to enhance the amount and quality of the information extracted from the data analyzed given the human capacity to adapt to the data. The authors hope the latter will enhance the effectiveness of already developed sonification techniques and the improvements to the interface with the sonification prototype developed. In our exploration this is done with non linear/ non-stationary data.

Due to the intrinsic plasma nonlinearities, its acceleration process (accelerated charged particles precipitating along magnetic field lines) and its associated phenomena, such as the generation of magnetic fields, are notoriously difficult to unravel. The base of expansion for non-stationary and non-linear data has to be adaptive [Huang et al. 2006]. That means the definition of the basis has to be data dependent, meaning a posteriori-defined basis, which implies an approach very different from the regular data analysis techniques in which the basis of expansion is established at the beginning [Bendat et al. 1990]. The human ability to adapt to the data and distinguish the statistical and dynamical characteristics of time-series data may lead to a more reliable and empirical data dependant analysis. The disadvantage of the sonification technique is its difficulty of laying a firm theoretical foundationfuture papers the authors hope to divulger results laying a firm theoretical foundation (a topic of continuing research). In the following pages we explore the use of a mple new sonification technique to do the first approach to magnetic field fluctuation data to establish, a posteriori, the numerical basis of expansion.  The sonification technique used has been developed to analyze crossing data from constellation of satellites, radio data etc to mention a few.

The magnetic parameters of interest are a strongly solar cycle dependent characteristic. We propose to sonify the power spectra of the magnetic fluctuation data to analyze the frequency content of the data. The registry space spanned (the relative high or low frequency of discrete or noise sound) would give information on instantaneous frequency changes. The notes are approached as the product of several frequencies (harmonics, in-harmonics, fundamentals, etc). Then magnetic temporal fluctuation information is portrayed as a simultaneously sounded cluster of pitches (a,b,c,d,e,f,g,…etc). Each data set may be sonified/mapped on pitch active for phase, frequency and time variations.

Conclusions

The application of ‘handy’ sonification techniques ( Easy to use and easy to explain , Barras et. al) may be a useful tool to analyze instantaneous frequency changes in the data. These ‘handy’ sonification techniques may help scientists not to miss important but not obvious changes in the non linear, non stationary data like astrophysics data. Rigorous experiments in perception psychology will help to improve and enhance the information/analysis of non-linear, non stationary astrophysics data using sonification techniques.

The sonification technique proposed in this paper has to be validated further, together with an usability evaluation for usefulness, satisfaction and effectiveness. Further validation gathering data from a constellation of satellites may validate/improve further the sonifications applied in this research: GOES, THEMIS and WIND.

The frequencies heard in the sonification may be associated with astrophysical plasmas like the solar wind. If a cavity is excited, then a stellar outflow may be associated to the changes heard. If it is so then characteristic size can be roughly estimated using f as 46 mHz excitation at c/f = 6.5E9 meters or 9.3 solar radii. An outflow is further suggested by the presence of a turbulent plasma listening for spectral index will corroborate the latter.

References

Chen, Q., N. E. Huang, S. Riemenschneider, and Y. Xu, 2005: A B-spline approach for empirical mode decomposition. Adv. Comput. Math. in press.

Cohen, L., Time-Frequency Analysis. Prentice Hall, 299,1995.

Diks, C., Nonlinear Time Series Analysis: Methods and Applications. World

Scientific Press, 180, 1999.

GrÅNochenig, K., 2001: Foundations of Time-Frequency Analysis. BirkhÅNauser359, 2001.

Hahn, S.,Hilbert Transforms in Signal Processing. Artech House 442, 1995.

Harrar L., and Stockman T. 2007. Designing Auditory Graphs Overviews: An Examination Of Discrete vs Continuous Sound And The Influence Of Presentation Speed. ICAD 2007.

Kantz, H., and T. Schreiber, 1999: Nonlinear Time Series Analysis. Cambridge University Press, 304,1999.

Hermann T., Meinicke P., and Ritter H.. Principal Curve Sonification. ICAD 2002

Lunney, D and Morrison R. Auditory Presentation of Experimental Data. SPIE 259, 40-46,(1990).