I have published 2 semantic web resources developed during different phases of the project. I have tried to follow the best practice recipes for publishing data in RDF format and used the Vapour linked data validator to ensure the data is published according to best practices recommendations.
The first resource is a catalog of audio features compiled from journal articles, feature extraction software source code, online documentation and existing linked data resources. The catalog does not have much structure and contains information mostly in text format. It provides information about the domain of audio features and serves as the starting point for the ontology engineering process. The catalog can be accessed here:
Based on this catalog, we created a vocabulary of audio features which provides terms for the modular audio feature ontology framework. It does not seem feasible that a single ontology completely represents the different conceptualisations of the domain that exist in audio research. For example, in some contexts, audio features are categorised according to psychoacoustic concepts, such as pitch, rhythm and timbre, while in others the computational workflow used in calculating the features determines their hierarchical organisation. Therefore we have opted to create three levels for the framework:
- high level ontology representing the domain in abstract terms
- a flat vocabulary listing all the features
- modular tool and task specific ontologies that can be built using the ontology and the vocabulary
The vocabulary is available here:
The high level ontology and examples of tool specific ontologies will be published on this website soon.