Javascript must be enabled for the correct page display

Development of a tool for Soundscape Annotation - What do we hear when we listen?

Linden, R. van der (2011) Development of a tool for Soundscape Annotation - What do we hear when we listen? Master's Thesis / Essay, Artificial Intelligence.

[img]
Preview
Text
thesis_rvdlinden_review22dec2011.pdf - Published Version

Download (8MB) | Preview
[img] Text
AkkoordvanderLinden.pdf - Other
Restricted to Repository staff only

Download (49kB)

Abstract

In the developing field of automatic sound recognition there exists a need for well-annotated training data. These data currently can only be gathered through manual annotation; a time-consuming and sometimes tedious task. How can a software tool support this task? The objective of this master’s project is to develop and validate a tool for soundscape annotation. Furthermore we assess the strategies that subjects employ when annotationing a real-world sound recording. In an experiment with untrained participants, annotations were collected together with user data (keystrokes and mouse clicks) that provide insight in the strategies subjects employ to achieve the annotation task. Dividing attentional resources over the time span of the recording is an impor tant aspect of the task. Soundscape annotation, the process of annotating a real-world sound recording, can be seen as a special case of ’everyday listening’ (Gaver). When annotating an audio recording offline (as opposed to reporting auditory events ’in vivo’) the subject lacks context knowledge, but offline annotation also opens new possibilities for the listener, for example to listen to the same sound event more than once. These differences have implications for the task and ultimately bring the question to mind: what makes a ’good’ annotation?

Item Type: Thesis (Master's Thesis / Essay)
Degree programme: Artificial Intelligence
Thesis type: Master's Thesis / Essay
Language: English
Date Deposited: 15 Feb 2018 07:46
Last Modified: 15 Feb 2018 07:46
URI: http://fse.studenttheses.ub.rug.nl/id/eprint/9834

Actions (login required)

View Item View Item