Audio Features Ontology Specification

This ontology expresses some common concepts to represent some features of audio signals. It mainly relies on the Event ontology, in order to classify particular parts of the timeline backing an audio signal. It also supports dense features, such as chromagrams, onset detection function etc.

Classes: | Amplitude | Beat | ChordSegment | Chromagram | DetectionFunction | EmotionSegment | EnglishIdiom | GermanIdiom | HungarianIdiom | Idiom | IdiomSegment | KeyChange | KeySegment | Laugh | LinearFrequencyCentroid | LogFrequencyCentroid | Loudness | Major | Minor | ModeChange | MusicSegment | MusicStructuralSegment | NonTonalOnset | Onset | OnsetDetectionFunction | Ornament | PersonSpeaking | Pitch | Point | Segment | Signal | SpectralCentroid | SpeechSegment | StructuralSegment | Tempo | TempoDetectionFunction | Text | TimeSignature | TonalChange | TonalChangeDetectionFunction | TonalContentSpace | TonalOnset | TonicChange | TonicSegment | ZeroCrossing | ZeroCrossingCount |

Properties: | confidence | dimensions | emotional_intensity | feature | idiom | key | new_key | new_mode | new_tonic | person | pitch | signal_feature | tempo | text | tonic | value |

Individuals: | austrian_german | english_irish_accent | english_scottish_accent | standard_british_english | standard_german | standard_hungarian | standard_us_english | swiss_german |

Classes and Properties (full detail)

Class: af:Amplitude - testing -

Amplitude - Result of an amplitude following process

[back to top]


sub-class-of:Signal

Class: af:Beat - testing -

Beat - A beat event (instantaneous)

[back to top]


sub-class-of:Point

Class: af:ChordSegment - testing -

ChordSegment - A classifier for chords. Equivalent concept as the one
in the Chord ontology.

[back to top]


sub-class-of:MusicSegment

Class: af:Chromagram - testing -

Chromagram - A chromagram feature.

[back to top]


sub-class-of:Signal

Class: af:DetectionFunction - testing -

DetectionFunction - A detection function.

[back to top]


sub-class-of:Signal

Class: af:EmotionSegment - testing -

EmotionSegment - A classifier for emotional content

[back to top]


in-domain-of:emotional_intensity
sub-class-of:SpeechSegment

Class: af:EnglishIdiom - testing -

EnglishIdiom - English idiom

[back to top]


Class: af:GermanIdiom - testing -

GermanIdiom - German idiom

[back to top]


Class: af:HungarianIdiom - testing -

HungarianIdiom - Hungarian idiom

[back to top]


Class: af:Idiom - testing -

Idiom - Idiom (English with Irish accent, Austrian German, etc.)

[back to top]


in-range-of:idiom

Class: af:IdiomSegment - testing -

IdiomSegment - A classifier for recognized idiom (English with Irish accent, etc.)

[back to top]


in-domain-of:idiom
sub-class-of:SpeechSegment

Class: af:KeyChange - testing -

KeyChange - A key change event. The factor of such an event captures the key that holds after that event.

[back to top]


in-domain-of:new_key
sub-class-of:Point

Class: af:KeySegment - testing -

KeySegment - A classifier for keys.
In case of a "clean cut", instances of such events
should have one factor: the detected key (perhaps using the
key ontology at http://purl.org/NET/c4dm/keys.owl).
In other cases, we can put several keys as a factor, maybe with different confidence
factors, using event decomposition.

[back to top]


in-domain-of:key
sub-class-of:MusicSegment

Class: af:Laugh - testing -

Laugh - A classifier for laugh content

[back to top]


sub-class-of:SpeechSegment

Class: af:LinearFrequencyCentroid - testing -

LinearFrequencyCentroid - Log-frequency spectral centroid

[back to top]


sub-class-of:SpectralCentroid

Class: af:Loudness - testing -

Loudness - Event holding a loudness value

[back to top]


sub-class-of:MusicSegment

Class: af:Major - testing -

Major - A classifier for a major mode region.
In case of a "clean cut", instances of such events have one
factor. In other cases, this event can have several factors, each associated
with a particular confidence using event decomposition.

[back to top]


sub-class-of:MusicSegment

Class: af:Minor - testing -

Minor - A classifier for a minor mode region.
In case of a "clean cut", instances of such events have one
factor. In other cases, this event can have several factors, each associated
with a particular confidence using event decomposition.

[back to top]


sub-class-of:MusicSegment

Class: af:ModeChange - testing -

ModeChange - A mode change event. The factors of such events include the mode that holds after the event.

[back to top]


sub-class-of:Point

Class: af:MusicSegment - testing -

MusicSegment - A classifier capturing the notion of an audio segment holding music.
This classifier can be subsumed with more specific classifiers.

[back to top]


sub-class-of:StructuralSegment

Class: af:MusicStructuralSegment - testing -

MusicStructuralSegment - Classifier capturing the notion of chorus, verse, intro, bridge, phrase, etc.

[back to top]


sub-class-of:MusicSegment

Class: af:NonTonalOnset - testing -

NonTonalOnset - A percussive onset

[back to top]


sub-class-of:Onset

Class: af:Onset - testing -

Onset - A classifier corresponding to the output of an onset detection process.

[back to top]


sub-class-of:Point

Class: af:OnsetDetectionFunction - testing -

OnsetDetectionFunction - An onset detection function

[back to top]


sub-class-of:DetectionFunction

Class: af:Ornament - testing -

Ornament - Classifying a temporal region holding a musical ornament

[back to top]


sub-class-of:MusicSegment

Class: af:PersonSpeaking - testing -

PersonSpeaking - A classifier associating a FOAF description to a particular segment
(to classify genre, recognized speaker, etc.)

[back to top]


sub-class-of:SpeechSegment

Class: af:Pitch - testing -

Pitch - Associate a point to a particular pitch

[back to top]


in-domain-of:pitch
sub-class-of:Point MusicSegment

Class: af:Pitch - testing -

Pitch - Event holding a pitch value (for example coming from a monophonic transcription process)

[back to top]


in-domain-of:pitch
sub-class-of:Point MusicSegment

Class: af:Point - testing -

Point - A classifier for instants corresponding to the output of an automatic process---an
artificial cognitive agent (should be associated to a time:Instant---DL version).

[back to top]


sub-class-of:event:Event

Class: af:Segment - testing -

Segment - A classifier for a temporal region corresponding to the output
of an automatic process---an artificial cognitive agent.

[back to top]


in-domain-of:feature
sub-class-of:event:Event

Class: af:Signal - testing -

Signal - A signal-like feature, holding dense data describing another signal.
Examples of signal features include chromagrams, spectrograms, onset detection functions etc.

[back to top]


in-range-of:signal_feature
in-domain-of:dimensions value
sub-class-of:mo:Signal

Class: af:SpectralCentroid - testing -

SpectralCentroid - Spectral centroid

[back to top]


sub-class-of:Signal

Class: af:SpeechSegment - testing -

SpeechSegment - A classifier capturing the notion of an audio segment holding speech
content

[back to top]


sub-class-of:StructuralSegment

Class: af:StructuralSegment - testing -

StructuralSegment - A classifier trying to capture the notion of structure in an audio piece.
This classifier should be subsumed by more specific classifiers: speech/music
segmentation, structural music segmmentation (intro, verse, chorus, etc.).

[back to top]


sub-class-of:Segment

Class: af:Tempo - testing -

Tempo - Event holding a tempo value (120bpm...)

[back to top]


in-domain-of:tempo
sub-class-of:MusicSegment

Class: af:TempoDetectionFunction - testing -

TempoDetectionFunction - A tempo detection function

[back to top]


sub-class-of:DetectionFunction

Class: af:Text - testing -

Text - A classifier allowing to associate some text to a segment

[back to top]


in-domain-of:text
sub-class-of:SpeechSegment

Class: af:TimeSignature - testing -

TimeSignature - Classifying a temporal region with a particular time signature (what? there is something else than 4/4?? :-)

[back to top]


sub-class-of:MusicSegment

Class: af:TonalChange - testing -

TonalChange - A tonal change event.

[back to top]


sub-class-of:Point

Class: af:TonalChangeDetectionFunction - testing -

TonalChangeDetectionFunction - A tonal change detection function

[back to top]


sub-class-of:DetectionFunction

Class: af:TonalContentSpace - testing -

TonalContentSpace - A 6-D tonal content space

[back to top]


sub-class-of:Signal

Class: af:TonalOnset - testing -

TonalOnset - A pitched onset

[back to top]


sub-class-of:Onset

Class: af:TonicChange - testing -

TonicChange - A tonic change event. The factors of such events include the tonic that holds after the event.

[back to top]


sub-class-of:Point

Class: af:TonicSegment - testing -

TonicSegment - A classifier for tonics.
In case of a "clean cut", instances of such events have one
factor. In other cases, this event can have several factors, each associated
with a particular confidence using event decomposition.

[back to top]


in-domain-of:tonic
sub-class-of:MusicSegment

Class: af:ZeroCrossing - testing -

ZeroCrossing - Classifier for a zero-crossing point

[back to top]


sub-class-of:Point

Class: af:ZeroCrossingCount - testing -

ZeroCrossingCount - Zero-crossing counts

[back to top]


sub-class-of:Signal

Property: af:confidence - testing -

confidence - A really generic property, allowing to associate a confidence to a document or to a particular
resource (such as an automatically detected segment)

Domain:
Range: rdfs:Literal

Property: af:dimensions - testing -

dimensions - Associates a signal with its dimensions, eg. "12 3546". Equivalent to the mpeg7:dim attribute of the mpeg7:Raw
element.

Domain: Signal
Range:

Property: af:emotional_intensity - testing -

emotional_intensity - Associates an "emotional intensity" value to a
segment detected as being "emotional"

Domain: EmotionSegment
Range: rdfs:Literal

Property: af:feature - testing -

feature - Associates a segment classifying a feature to the actual feature---we should use mpeg7 sort-of datatypes here

Domain: Segment
Range: rdfs:Literal
Sub-property-of: event:literal_factor

Property: af:idiom - testing -

idiom - Associates a segment which classify an idiom (English with Irish accent, etc.) to an actual idiom

Domain: IdiomSegment
Range: Idiom
Sub-property-of: event:hasLiteralFactor

Property: af:key - testing -

key - Associates a key event to the corresponding key
Domain: KeySegment
Range: owl:Thing
Sub-property-of: event:factor

Property: af:new_key - testing -

new_key - Associates a key change event to the new key, holding after the change
Domain: KeyChange
Range: owl:Thing
Sub-property-of: event:factor

Property: af:new_mode - testing -

new_mode - Mode that holds after the mode change event
Domain: af:ModeChangeEvent
Range: owl:Thing
Sub-property-of: event:factor

Property: af:new_tonic - testing -

new_tonic - Tonic that holds after the tonic change event
Domain: af:TonicChangeEvent
Range: owl:Thing
Sub-property-of: event:factor

Property: af:person - testing -

person - Associates a segment classified as holding speech content to a FOAF
profile (maybe holding information about genre, recognized speaker URI, etc.)

Domain: af:PersonSegment
Range: foaf:Person

Property: af:pitch - testing -

pitch - Associates a pitch event to the corresponding pitch

Domain: Pitch
Range:
Sub-property-of: feature feature

Property: af:signal_feature - testing -

signal_feature - Associates an audio signal with a track-level signal feature.
eg. a spectrogram, a chromagram, an onset detection function.

Domain: mo:Signal
Range: Signal

Property: af:tempo - testing -

tempo - Associates a tempo event to the corresponding tempo value.

Domain: Tempo
Range:
Sub-property-of: feature feature

Property: af:text - testing -

text - Relates a Text segment (classifying a time interval that can
be associated to some text) to a text string or other literals (number, etc.)

Domain: Text
Range: rdfs:Literal
Sub-property-of: event:hasLiteralFactor

Property: af:tonic - testing -

tonic - Associates a tonic event with the corresponding tonic
Domain: TonicSegment
Range: owl:Thing
Sub-property-of: event:factor

Property: af:value - testing -

value - Associates a signal resource as defined in the Music Ontology to a literal: its actual value.
For example, it can be used to link an onset detection function to a list of values.

Domain: Signal
Range:
Sub-property-of: rdf:value

Individual: austrian_german -

Austrian German -
Class: af:GermanIdiom

Individual: english_irish_accent -

English with an Irish accent -
Class: af:EnglishIdiom

Individual: english_scottish_accent -

English with a Scottish accent -
Class: af:EnglishIdiom

Individual: standard_british_english -

Standard British English -
Class: af:EnglishIdiom

Individual: standard_german -

Standard German -
Class: af:GermanIdiom

Individual: standard_hungarian -

Standard Hungarian -
Class: af:HungarianIdiom

Individual: standard_us_english -

Standard US English -
Class: af:EnglishIdiom

Individual: swiss_german -

Swiss German -
Class: af:GermanIdiom