Responsive image

BAT: An open-source, web-based audio events annotation tool

Blai Meléndez-Catalán, Emilio Molina, Emilia Gómez
In this paper we present BAT (BMAT Annotation Tool), an open-source, web-based tool for the manual annotation of events in audio recordings developed at BMAT (Barcelona Music and Audio Technologies1). The main feature of the tool is that it provides an easy way to annotate the salience of simultaneous sound sources. Additionally, it allows to define multiple ontologies to adapt to multiple tasks and offers the possibility to cross-annotate audio data. Moreover, it is easy to install and deploy on servers. We carry out an evaluation where 3 annotators use BAT to annotate a small dataset composed of broadcast media recordings. The results of the experiments show that BAT offers fast annotation mechanisms and a method to assign salience that produces high agreement among annotators.
            
@inproceedings{2017_EA_17,
  abstract = {In this paper we present BAT (BMAT Annotation Tool), an open-source, web-based tool for the manual annotation of events in audio recordings developed at BMAT (Barcelona Music and Audio Technologies1). The main feature of the tool is that it provides an easy way to annotate the salience of simultaneous sound sources. Additionally, it allows to define multiple ontologies to adapt to multiple tasks and offers the possibility to cross-annotate audio data. Moreover, it is easy to install and deploy on servers. We carry out an evaluation where 3 annotators use BAT to annotate a small dataset composed of broadcast media recordings. The results of the experiments show that BAT offers fast annotation mechanisms and a method to assign salience that produces high agreement among annotators.},
  address = {London},
  author = {Meléndez-Catalán, Blai and Molina, Emilio and Gómez, Emilia},
  booktitle = {Proceedings of the International Web Audio Conference},
  editor = {Thalmann, Florian and Ewert, Sebastian},
  month = {August},
  pages = {},
  publisher = {Queen Mary University of London},
  series = {WAC '17},
  title = {BAT: An open-source, web-based audio events annotation tool},
  year = {2017},
  ISSN = {2663-5844}
}