Assignment 2: Audio Signal Processing for Data Sonification

DECO1013: Sound Design and Sonification

AYOU8415 - 310192668

Studyingthe differing effectiveness of information presentation through various senses, it quickly becomes apparent that patterns and the presence of inconsistency in a set of data are best recognised aurally, as in music. Taking advantage of this comparatively acute sense, I have designed a system which sonifies the changes in a comparative dataset for the optimal consumption of a user. Specifically, the package will provide an audible representation of the relationship between two sets of data over time, producing a harmony of frequencies that is modulated in its pitch and composure by the given data.The ability to distinguish tonal impurity is common to most people, as such this package seeks to provide an experience which exercises that ability and better communicates the characteristics of a comparative dataset, complimentary to visual representation through the shift in pitch and fidelity of a musical harmony.

The designed sonification system works through the analysis of a comparative dataset. In the case of Google, this would be implemented within the Google Trends service, which generates such data in the analysis of search entries provided by users. In an end-user scenario, what this would mean is an audio file, produced by the sonification package, accompanying the existing visual infrastructure Google has put in place for the Google Trends service, at the user’s request. Technically speaking, the data is fed through the package, which creates a musical harmony based on the initial values. As the system reads through the data, the sound is modulated according to the changes in the relationship between the two (or more) subjects; greater difference in tone, creating notable dissonance, is achieved when the difference between datasets grows.

Thus far, theoretical tests on the design of the system have proved to be successful, although a functional practical application remains to be ascertained. When simplified to a format which is commonplace in the user’s cognitive space, that of music, it does in fact become easier for a user to grasp the relationship between two sets of otherwise overwhelmingly numerical data and as such, warrant the need for this system. Interestingly, as has been found in testing, while presenting data comparisons exclusively through visual means often gives an adequate representation of the general picture, the distribution of data sequentially and over time (in a sound file) appears to give users a better understanding of the dynamics of the dataset. It is the aim of this design to capitalise on this advantage while using musically based parameters to create a more familiar experience while still being just as informative.