The Audio Commons Ontology has been created in the context of the European project [Audio Commons](https://www.audiocommons.org/), which were aimed at favouring sharing and reuse of audio content (both of musical and non-musical nature), specially using [Creative Commons](https://creativecommons.org/) licenses. It is designed specifically to integrate access to multiple heterogeneously structured audio content providers. In particular the following ones have been considered in the project: - [Freesound](https://freesound.org/), a collaborative database of Creative Commons Licensed sounds; - [Jamendo](https://www.jamendo.com/), a music website for independent artists; - [Europeana Sounds](https://www.eusounds.eu/), a European project collecting and sharing data from several archives of sounds and sound-related media; - [Internet Archive](https://archive.org/), a non-profit library of millions of freely accessible media contents. ### The base model <center> ![A diagram showing the entities Work, Expression, Manifestation, Item](./img/FRBR-Group-1-entities-and-basic-relations.svg.png) </center> The Audio Commons ontology is based on [Functional Requirements for Bibliographic Records (FRBR)](https://en.wikipedia.org/wiki/Functional_Requirements_for_Bibliographic_Records), a bibliographic model designed for user tasks of retrieval and access in online library catalogues and bibliographic databases. FRBR distinguish the products of intellectual or artistic endeavor in four main classes: work, expression, manifestation, item. It is available as RDF vocabulary at https://vocab.org/frbr/core. For further information on FRBR, check the [standard](https://repository.ifla.org/handle/123456789/811). Another important model related to audio is the [Music Ontology](http://musicontology.com/), which is also based on the FRBR model. Our ontology is more generic than the Music Ontology but is interoperable with it to support the specificities of music-related audio content. The Audio Commons ontology works also in conjunction with - [Dublin Core](https://www.dublincore.org/specifications/dublin-core/dcmi-terms/) for basic metadata, - [EBU Core](https://www.ebu.ch/metadata/ontologies/ebucore/) for technical properties of audio files, - [SKOS](https://www.w3.org/TR/skos-primer/) to support taxonomic classifications, - [Creative Commons Licensing schema](https://creativecommons.org/ns) to represent media licensing, - [Event ontology](http://motools.sf.net/event/event.html) for the formalisation of events, - [Audio Features Ontology](https://w3id.org/afo/onto/) for the characterisation of audio features. ### Ontology overview ![A diagram showing the main classes of the Audio Commons ontology](./img/AC-Ontology-Skeleton.png) The figure above shows the most general classes and properties of the Audio Commons ontology and their relationship with classes of the FRBR and the Music Ontology: - **`ac:AudioExpression`**, the specific intellectual or artistic form that a work takes each time it is _realised_, in the audio domain, e.g., the recording or synthesis of music or sounds; - **`ac:AudioManifestation`**, the physical embodiment of an audio expression, e.g., a musical track, a sound, an album; - **`ac:AudioItem`**, a single exemplar of an audio manifestation, e.g., a copy of a CD or a specific media file. The FRBR class Work, representing a distinct intellectual or artistic creation on a more conceptual level, has not been specialised in Audio Commons. The Audio Commons ontology provides a generic schema for audio classification through the property **`ac:audioCategory`** that associates any audio expression, manifestation, or item to some generic **`ac:AudioCategory`**. These two terms can be specialised to provide spefic categorisations. ### Manifestations: audio clips and collections ![A diagram showing the main classes of the Audio Commons ontology](./img/AC-Ontology-Manifestation.png) The figure above shows classes and properties related to audio manifestations. An instance of **`ac:AudioClip`**, a subclass of `ac:AudioManifestation`, is any audio segment that has been published in some form or uploaded for consumption, for example, a track in a music label's repository or a sound in an audio repository, library or archive. The class **`ac:AudioCollection`**, another subclass of `ac:AudioManifestation`, is used to represent collections of audio clips. The class **`ac:AudioCollectionNode`** is used to represent single nodes of a collection, offering local information like the index in the collection and pointers to previous and next nodes. The separation between the collection node and its actual content (e.g., an `ac:AudioClip`) permits the same content to be shared in multiple collections. The content of each node of a collection is not limited to an `ac:AudioClip`, but may contain any `ac:AudioManifestation`. Collections can thus contain other collections to support specific cases, e.g. a mapping to the Music Ontology model where an `mo:Release` can contain multiple `mo:Record`(s) that can in turn contain multiple `mo:Track`(s). ### Items: audio files ![A diagram showing ac:Item class and related class and properties](./img/AC-Ontology-Item.png) An `ac:AudioItem` is a concrete exemplar of an audio manifestation. In our domain, the main exemplars are the actual audio files. The corresponding class **`ac:AudioFile`** is a subclass of `ac:AudioItem`. It is subclass of `ebu:MediaResource` too and the corresponging properties of EBU Core can be used to describe the file (e.g., `ebu:hasEncodingFormat`, `ebu:fileSize`). ### Expressions: audio signals ![A diagram showing ac:Expression class and related class and properties](./img/AC-Ontology-Expression.png) While `ac:AudioFile` represents a concrete file encoded in a certain format, **`ac:DigitalSignal`** is the representation of the corresponding digital signal. `ac:DigitalSignal` is a subclass of `ac:MusicalExpression`. The data properties `ac:sampleRate`, `ac:bitsPerSample`, and `ac:channels`, associate a signal with its basic features specific to digital representations. The property `ac:publicationOf` can be used to associate an `ac:AudioClip` with the corresponding digital signal. The property `ac:encodes` instead, associates an `ac:AudioFile` with the encoded digital signal. ### Events: recording/syntesis ![A diagram showing ac:SignalProduction class and related class and properties!](./img/AC-Ontology-SignalProduction.png) The description of temporal events is crucial to formalise and document transitions in the workflow of audio production and publication. The class `event:Event` of the Event Ontology is thus specialised for specific actions that are interesting for the audio domain. Using the `event:Event` class, details of an event such as its location in time and space, its factor and its products may be explicitly described. Moreover, events can be composed using the property `event:sub_event`, to build complex events. The class `ac:SignalProduction` represents the act of producing a `ac:Signal`, which could be either an `ac:AnalogSignal` or a `ac:DigitalSignal`. This is complemented by **`ac:Recording`**, representing the process of recording a sound (`ac:Sound`) or the product of **`ac:Synthesis`**, in case of artificially generated sounds. ### Events (continued): publication ![A diagram showing ac:Publication class and related class and properties!](./img/AC-Ontology-Publication.png) The event **`ac:AudioPublication`** represents the public release of a piece of work, e.g., the release of a new album by a band. ### Usage Example ![A diagram showing an example of ontology usage!](./img/AC_Example-AudioClip.png) The figure above shows an example of use of the Audio Commons ontology to represent a sound from the Freesound database. For simplicity only the classes `ac:AudioClip`, `ac:AudioFile`, and `ac:AudioCategory` are used in this case. 2017/07/27 11:09:00 2017/07/27 11:09:00 The Audio Commons Ontology provides classes and properties for describing audio content, both musical and non-musical, on the Semantic Web. It is a specialisation of a bibliographic model called Functional Requirements for Bibliographic Records (FRBR). Furthermore, it contains the generalisation of some concepts from the Music Ontology. The Audio Commons Ontology ac https://w3id.org/ac-ontology/aco# The Audio Commons Ontology provides classes and properties for describing audio content, both musical and non-musical, on the Semantic Web. It is a specialisation of a bibliographic model called [Functional Requirements for Bibliographic Records (FRBR)](https://en.wikipedia.org/wiki/Functional_Requirements_for_Bibliographic_Records). Furthermore, it contains the generalisation of some concepts from the [Music Ontology](http://musicontology.com/). Revision: 1.3.0 An analog signal. Analog signal Classify some audio-related thing in a category. is in audio category Any category that is used to classify audio content. Audio category An audio segment that has been somehow published or uploaded for consumption (e.g., a track in a music label's repository or a sound in an audio repository, library or archive). Audio clip A collection of audio manifestations (e.g., a sound pack, an album, a result set of a search over a repository of audio content). As an audio collection is an audio manifestation too, collections may be nested. An audio collection is composed of nodes. Audio collection 1 1 A node that is the first in the containing audio collection First node 1 A node that is the last in the containing audio collection Last node 1 rdf:nil A node of an audio collection. It encapsulates the actual content (ac:nodeContent), the position in the collection (ac:nodeIndex), and links to the previous and following node in the collection (ac:previousNode and ac:nextNode). Node 1 1 1 1 A node that is not the first in the containing audio collection Any node but the first 1 A node that is not the last in the containing audio collection Any node but the last 1 The specific intellectual or artistic form that a work takes each time it is “realized”, in the audio domain (e.g., the recording or synthesis of music or sounds). Audio expression Associates a signal to an extracted audio feature. audio feature A concrete audio file (encoded in some format) Audio file A concrete exemplar of an audio manifestation (e.g., a copy of a CD or a specific media file). Audio item The physical embodiment of an audio expression (e.g., a musical track, a sound, an album). Audio manifestation MD5 checksum computed on the audio content of the file. audio MD5 checksum The act of publishing some audio content as an audio manifestation. Audio publication Used to relate a person or a group of people who authored an audio manifestation. author Relates an audio manifestation to an audio item (e.g., a media file). available as Relates an audio item (e.g., a media file) to the corresponding audio manifestation. available item of Associates a digital signal to the number of bits used to encode one sample. bits per sample Associates a signal to the number of channels it holds (mono → 1, stereo → 2). channels Convenience alias of ac:firstNode to attache an ac:AudioCollection to an rdf:List that holds the members collection as list Used to relate a person or a group of people who compiled an audio manifestation. compiled Associates an audio manifestation with the collection nodes that contain it. container node Associates an audio manifestation with a node in the default parent collection (e.g., a musical track to the album it has been published it), if there is such a thing. default container node A digital signal. Digital signal The duration of an audio clip or a signal in ms. duration Associates a MusicalItem (a track on a particular CD, an audio file, a stream somewhere) to the signal it encodes. encodes Associates an audio collection to its first node. first node Links a sound, an artist, a record, etc. to a corresponding webpage. homepage Indicates a pictorial image (JPEG, GIF, PNG, Etc.) associated to a sound, an artist, a record, etc. image Associates a signal to a MusicalItem (a track on a particular CD, an audio file, a stream somewhere) that encodes it. is encoded by Associates an audio collection to its last node. last node Associates an audio collection to its nodes. member node Associates a node with the next node in the collection. next node Associates a node with the actual content (an audio manifestation). node content Associates a node with its position in the collection. node index Relates an audio clip to the original published audio file (if there is such thing). original file Associates a node with the previous node in the collection. previous node Associates a signal production event to the produced signal. produced signal Associates a signal production act to a sound from which the signal was produced (e.g., if recording, the recorded sound). production from Link a particular manifestation to the related expression (e.g., a signal or a sound). publication of Used to relate a person or a group of people who published an audio manifestation. published Link an expression to its publication as manifestation. published as Associates a publishing act to the audio manifestation. published audio manifestation The act of recording some sound to produce a signal. Recording Associates a recording act to the recorded (physical) sound. recording of Associates a digital signal to its sample rate. sample rate A subclass of AudioExpression, representing a signal, for example a master signal produced by a performance and a recording. Signal Associates the signal encoded by an audio item to an extracted audio feature. Shortcut for ac:encodes/ac:audioFeature 1 signal audio feature The number of channels the digital signal holds (mono → 1, stereo → 2). Shortcut for ac:encodes/ac:channels 1 signal channels The duration of the signal in ms. Shortcut for ac:encodes/ac:duration 1 signal duration The act of producing a signal (e.g., through a recording or a synthesis). Signal production The sample rate of the digital signal. Shortcut for ac:encodes/ac:sampleRate 1 signal sample rate A subclass of AudioExpression, representing a (physical) sound. Physical sound The act of synthesizing a signal. Syntesis George Fazekas George Fazekas Miguel Ceriani Miguel Ceriani