Other Free Encyclopedias » Online Encyclopedia » Encyclopedia - Featured Articles » Contributed Topics from A-E

Context-Aware Musical Audio

information piece related feature

Definition: Context-aware musical audio is the association of any information related to the piece of music.

The phrase context-aware multimedia , in the case of audio documents in the music domain, can be defined as the association of any information related to the piece of music extending the information required for or supporting playback. Furthermore, it could be understood as the usage of this information in application scenarios by Music Information Retrieval (MIR) systems, in order to create context-aware multimedia applications.

Such information can either be found implicitly in the content of a piece of music (often referred to as content-based or low-level features) or is associated with the audio instances as external metadata (high-level features). Content-based information is extracted via feature extraction algorithms. Those algorithms are often complex and computational intensive, and one must consider the different formats in which a piece of music can be digitally stored. On the other hand, the symbolic representation of musical works allows easier access to various feature parameters such as pitch or rhythm-related attributes – even for polyphonic music. Musical pieces represented by samples, which will be encountered in most cases, raise the need to compute the symbolic representation out of the audio signal via signal processing techniques or to compute attributes that are not based upon symbolic music representation.

External metadata, understood as information which cannot be extracted out of the audio content of a musical piece itself, is the enabler for the next step in building context-awareness through interpreting or processing this metadata with the intention to retrieve or compute further information describing the piece of music. Such information is generally annotated manually and would be for example the artist, year of production, or music label. These attributes can be processed further to achieve results similar to those obtained when using implicit characteristics or used in combination with low-level features to improve search results, which can be a more efficient and faster method, depending on the application scenario.

Making use of the feature parameters mentioned above leads to context-aware multimedia applications. One example would be an audio player providing additional information related to the actual playback position in a piece of music, such as references to the score or lyrics. Another would be the recommendation of similar tracks related to a user’s favorite, based upon various feature parameters. Research related to this topic is very active, because the music industry is beginning to abandon the attitude of restricting digital distribution channels and is attempting instead to make use of them for their own profit. Research results will improve over time and one can expect increasing performance of feature-extraction algorithms, which will in turn provide a solid basis for context-aware multimedia applications in the music domain.

Context-Aware Video Production - Context Awareness, Context Enhanced Retrieval of Video, Vision [next] [back] Context-Aware Multimedia - Introduction, Context, Context-Awareness, Context-A ware Multimedia, Modeling, Retrieval, Authoring and Presentation, Summary

User Comments

Your email address will be altered so spam harvesting bots can't read it easily.
Hide my email completely instead?

Cancel or