Pie Chart Sonification Keith M. Franklin and Jonathan C. Roberts University of Kent, Computing laboratory, Canterbury, England, UK. {kmf1@kent.ac.uk, j.c.roberts@kent.ac.uk} Abstract Different acoustic variables such as pitch, volume, timbre and position can be used to represent quantitative, qualitative and categorical aspects of the information. Such sonifications are particularly useful for those with visual impairments; they are also beneficial in circumstances where visual representations would be impossible to use or to enrich a graphical realization. We demonstrate methods of representing an audible pie chart representation such that the hearer understands the information through an equivalent representation. We implement and evaluate five designs. In each the user is positioned at the center of the chart and perceives the information through positional sound sources. Keywords--- Sonification, Visualization, Charts. 1. Introduction Sound is very important in our every day life; often we use it without consciously knowing that we are. For example, when we walk down a meandering corridor we use sound made by the oncoming people to avoid collisions, alternatively at a music concert given by an amateur group we may particularly notice mistakes the amateurs have made, or if someone drops some coins then we may quickly perceive the value of the dropped coinage. Indeed, over the years sound has be used to represent more information-rich phenomenon, from representing errors by alarms (such as used in computer interfaces), through sorting algorithms [1] to more recently visualizing the web [2] or sonifying well-logs from oil and gas exploration [3]. Sonification may be readily used to allow the user to view patterns within the data. In particular, sound is especially important for people with visual impairments. Thus, it is important to find ways to represent information that is accessible to this community. There are other reasons why charts and diagrams should be represented by sound, in addition to making the representations accessible to partial or non-sighted users. For example, there may be situations where the user cannot view a screen because they are monitoring something else (e.g. in a machine room where the engineer is constantly monitoring the material being cut and machined), or sonification may be useful to represent information where only very small screens are available. Non-visual visualizations, particularly pie- chart sonification tender a range of particular challenges such as how to represent the concepts of two- dimensional charts into sound space, how to effectively map the values to sound, whether users can actually perceive the information and, in particular, how accurate do users perceive the information. In this paper, we present and evaluate some novel ways to represent pie charts using sound. Similarly to representing data by graphics (using the retinal variables of color, size, orientation, symbol etc) there are many sound attributes that can be used to represent the data. Thus, the challenge is how to use these variables to effectively map the data into audible parameters. We are particularly interested in non-speech mappings, certainly there are advantages in representing the information by speech, but such transformations loose spatial illustration and thus can miss-represent or allow the listener to miss- out on the richness of the underlying information. Also, some may argue that pie charts are not very good at effectively displaying the underlying data; but, pie charts are often used, appear in many texts and thus we believe should exist in an audible form. Obviously, a pie chart made from sound cannot be identical to its visual counterpart, but inspiration can come from the visual pie chart itself. This idea of generating an equivalent representation is supported by current teaching methodologies; for example, blind and visually impaired users feel visual representations of diagrams using swell paper [4]. Using this method they develop a similar mental model to normal sighted users. Thus, by creating a sonic pie chart that is based on the visual representation, users will be able to draw on their previous experience and knowledge of pie charts. 2. Background There are various papers that demonstrate chart sonification, for example, Ramloll [5] and Bonebright [6] represent line graphs and earlier work by Brown [1], when he represents the state of the sorting algorithms, could be classified as bar-chart sonification. However, we have to look at the haptic literature to find a non- visual pie chart representation. Indeed, there are a reasonable number of papers on visualizing charts through haptics. Many of the methods single point