We are happy to announce that after 9 months since the start of the AudioCommons project we have reached, on due time, our first milestone! In the following paragraphs we briefly summarise the work that we’ve carried out in different aspects of the project. Most of the documents that are produced in the project as well as scientific publications and other kinds of relevant material can be found in the materials section of our website. A broader introduction to the Audio Commons initiative can be found in our position paper.
During the last months we have looked into intellectual property and legal aspects related to the Audio Commons Ecosystem. We have studied the Creative Commons licensing framework and made recommendations on how to apply it in the Audio Commons context and in the audio context in general. We have looked at which license-related aspects should the technology behind the Audio Commons Ecosystem support, including what steps should be followed in ‘standardised’ (re)licensing processes for safe commercial reuse of content, and how should licenses be presented to users to maximise their understandability. The outcomes of this research can be fond in two documents, Report on Rights Management requirements and Report on usage of Creative Commons licenses, publicly available in our website.
In parallel to that, we started the design of the Audio Commons Ontology which will be used in the backbone of the Audio Commons Ecosystem. An ontology is a tool to model and conceptualize knowledge about a domain. The Audio Commons Ontology will therefore model knowledge about the sound and music domain such as what metadata fields are relevant to annotate different kids of audio resources, how can Creative Commons licenses be combined and interpreted, an other knowledge such as relations betweens music genres and music instruments. This will be a great resource to power intelligent sound and music systems connected to the Audio Commons Ecosystem. You can find more information about the initial design of the ontology in the Draft Ontology Specification document.
Finally, another other important work thread that we have been advancing on is the automatic annotation of musical and non-musical audio content. Of special relevance is our review of existing state of the art music annotation methods, and the work carried out for defining a hierarchy of timbral semantic descriptors aimed at describing non-musical audio concepts such as depth or brightness. The figure below shows a graphical representation of the hierarchy. An interactive version has also been made available. These documents will serve as a guide point for our future research in automatic sound and music annotation.
The next milestone is scheduled for the end of April 2017, and includes the release of a prototype implementation of the Audio Commons Ecosystem. We have already started working on the different software services that will conform it and on the definition of the common API that will allow 3rd party applications to interact with Audio Commons content. We will keep you updated!