While ‘Visualization’ focuses on visual representations of data, ‘Sonification’ focuses on the auditory equivalent.
Most sonifications used today (such as a car’s reverse proximity indication or a heart monitor) are based on synthetic audio, typically alternating between pure tones.
This project focuses on suggesting an alternative, by using natural audio in order to communicate the necessary information.
The algorithm developed searches a large repository of natural sounds using an efficient approximate-nearest-neighbour technique to find the best match to the data in the encoded version of the audio repository. A demonstration of the method is provided by communicating an artist’s popularity rate over time by finding a musical segment from his oeuvre that has a corresponding amplitude pattern.