definition

Sonification – A Definition

What is Sonification? This page gives a definition. A more detailed discussion of was published in my ICAD’ 08 article Taxonomy and definitions for sonification and auditory display (pdf, 340k). Despite the long existence of sonification in science, the existing definitions are rather fuzzy. The so far most commonly agreed definition of sonification is given in the Sonification Report: Status of the Field and Research Agenda (1999) that states:

Sonification is the use of nonspeech audio to convey information. More specifically, sonification is the transformation of data relations into perceived relations in an acoustic signal for the purposes of facilitating communication or interpretation.

Sound Artists and Musicians, who have been using data for compositional purposes for a long time, now start to denote their compositions as sonification, which raises the question what criteria need to be fulfilled for a sound to be called a sonification. Furthermore, the perceptual vicinity of sounds in modern electronic music to sounds of some sonifications, plus the fact that both involve data, raises the question to clarify important preconditions and requirements for sonification.

Definition: Sonification (by Thomas Hermann, ICAD 2008 paper (pdf, 340k))

Any technique that uses data as input, and generates (eventually only in response to additional excitation or triggering) sound signals may be called sonification, if and only if

  • (A) the sound reflects properties / relations in the input data.
  • (B) the transformation is completely systematic. This means that there is a precise definition of how interactions and data cause the sound to change.
  • (C) the sonification is reproducible: given the same data and identical interactions/triggers the resulting sound has to be structurally identical.
  • (D) the system can intentionally be used with different data, and also be used in repetition with the same data.

Definition: Sonification (short version)

Sonification is the data-dependent generation of sound, if the transformation is systematic, objective and reproducible, so that it can be used as scientific method.

(Thanks to Katharina Vogt for the productive discussion resulting in this 1-phrase definition.)

Discussion

General remarks:

  • According to this definition, the techniques Audification, Earcons, Auditory Icons, Parameter Mapping Sonification, Model-based Sonification are covered by the definition: They all represent information/data by using sound in an organized and well-structured way and they are thus different sonification techniques.
  • A distinction between data and information is – concerning the definition – irrelevant: information like for instance a ‘message’ can always be represented numerically and thus be understood as data. Whether sonification techniques aim at a more symbolic or analogic representation according to the taxonomy of Kramer is secondary for the definition.
  • Some articles have used “sonification” to refer specifically to mapping-based sonification, where data features are mapped to acoustic features of sound events or streams. In agreement with Greg Kramer, sonification is more generally the representation of data as sound. There may be times when a clear description (e.g. as model-based, or parameter-mapping) sonification may be helpful to avoid confusion with the general term of sonification.
  • Sonification refers to the technique and the process, the algorithm. Often and with equal right, the resulting sounds are called sonifications.
  • According to the definition, sonification is an accurate scientific method which leads to reproduceable results, addressing the ear rather then the eye (as visualization would do). Subjectivity in human interpretation is shared with other techniques that bridge the gap between data and the human sensory system.
  • Being already a scientific method, a prefix like in “scientific sonification” is not necessary.
  • Same as some data visualizations may be ‘viewed’ as art, sonifications may be heard as ‘music’, yet this use is different from their original intention.
  • Auditory Displays are systems that use sonification as technical (processing) subsystem, but in addition include the user, the interfaces (amplifier, loudspeaker, headphones) and respect the situational context (background noise, task, etc.)

Comments to (A):

  • Real-world acoustics are typically not a sonification (although they deliver object property specific systematic sound) since there is no external input data.
  • However, a bottle filling with rain, hit with a spoon once per second, can be seen as a sonification. The data here is the amount of rain, measured by the fill level, and the other conditions are also fulfilled.

Comments to (B):

  • Randomness may be allowed here (e.g. as temporal jitter to increase perceptability), yet it is important to declare where and what random elements are used (e.g. the mean, variance, distribution of used noise)

Comments to (C):

  • Sample-based identity is not necessary, yet all possible psychophysical tests should as a limit over many repetitions come to identical conclusions.

Comments to (D):

  • In result, “playing a musical instrument” is not a sonification of the performer’s emotional state, since it can not be repeated with the identical data. However, the resulting sound is a sonification of the interactions with the instrument (regarded as data), and in fact, music can be heard with the focus to understand interaction patterns.

Summary

The definition offers a clarification of the requirements for systems and methods to be called sonification. Potentially some requirements are missing, and I am eager to hear suggestions and problems that may help to further improve the definition.
Please send me an e-mail with any comments.

I thank Till Bovermann, Arne Wulf, Florian Grond, Alberto de Campo, Camille Peres, and in particular Gregory Kramer for the helpful discussions on the definition.

References