jSymbolic is a software application intended for conducting research in the fields of music information retrieval (MIR), music theory and musicology. Its primary purpose is to extract statistical information from musical data stored symbolically in file formats such as MIDI or MEI. This statistical information is formulated as feature values, which may be fed directly into automatic classification systems, may be used to query large musical datasets, or may be used by musicologists and music theorists for conducting empirical musical research.
jSymbolic includes an easy-to-use GUI, and may also be used via the command lien. It also has an API facilitating programmatic use. The software can be used either with its excellent general-purpose default settings, or advanced users can use it under a variety of settings (saved in a special configuration settings file).
Like all jMIR components, jSymbolic is free and open-source, and is designed to be used directly for conducting research as well as as a platform for iteratively developing new features that can then be shared amongst researchers. As such, jSymbolic emphasizes extensibility, and includes a modular design that facilitates the implementation and incorporation of new features, as well as the automatic provision of all other feature values to each new feature and dynamic feature extraction scheduling that automatically resolves feature dependencies. jSymbolic is implemented in Java in order to maximize cross-platform utilization.
jSymbolic is also part of the SIMSSA (Single Interface for Music Score Searching and Analysis) project, and is integrated with the music stored on the associated Elvis database.
jSymbolic is packaged with a library of 172 unique implemented features. Some of these are multidimensional, for a total of 1230 feature values.These features were developed through extensive analysis of publications in the fields of music theory, musicology and MIR. Most of these features had not previously been applied to MIR research, and many of them are entirely novel. The features can be loosely divided into the following seven categories:
- Pitch Statistics: How common are various pitches rel-ative to one another, in terms of both absolute pitches and pitch classes? How tonal is a piece? What is its range? How much variety in pitch is there?
- Melodic Intervals: What kinds of melodic intervals are present? How much melodic variation is there? What can be observed from melodic contour meas-urements? What types of phrases are used and how often are they repeated?
- Chords and Vertical Intervals: What vertical inter-vals are present? What types of chords do they repre-sent? How much harmonic movement is there, and how fast is it?
- Rhythm: Features are calculated based on the time intervals between note attacks and the durations of in-dividual notes. What meter and what rhythmic patterns are present? Is rubato used? How does rhythm vary between voices?
- Instrumentation: Which instruments are present, and which are emphasized relative to others? Both pitched and non-pitched instruments are considered.
- Texture: How many independent voices are there and how do they interact (e.g., polyphonic or homophon-ic)? What is the relative importance of voices?
- Dynamics: How loud are notes and what kinds of var-iations in dynamics occur?
Many more details on jSongMiner are available in the jSymbolic Manual.
McKay, C., T. Tenaglia, J. Cumming, and I. Fujinaga. 2017. Using statistical feature extraction to distinguish the styles of different composers. Accepted for publication at the Medieval and Renaissance Music Conference.
McKay, C., T. Tenaglia, and I. Fujinaga. 2016. jSymbolic2: Extracting features from symbolic music representations. Extended Abstracts for the Late-Breaking Demo Session of the 17th International Society for Music Information Retrieval Conference.
McKay, C. 2010. Automatic music classification with jMIR. Ph.D. Thesis. McGill University, Canada.
McKay, C., J. A. Burgoyne, J. Hockman, J. B. L. Smith, G. Vigliensoni, and I. Fujinaga. 2010. Evaluating the genre classification performance of lyrical features relative to audio, symbolic and cultural features. Proceedings of the International Society for Music Information Retrieval Conference. 213–8.
McKay, C., and I. Fujinaga. 2010. Improving automatic music classification performance by extracting features from different types of data. Proceedings of the ACM SIGMM International Conference on Multimedia Information Retrieval. 257–66.
McKay, C., and I. Fujinaga. 2008. Combining features extracted from audio, symbolic and cultural sources. Proceedings of the International Conference on Music Information Retrieval. 597–602.
McKay, C., and I. Fujinaga. 2007. Style-independent computer-assisted exploratory analysis of large music collections. Journal of Interdisciplinary Music Studies 1 (1): 63–85.
McKay, C., and I. Fujinaga. 2006. jSymbolic: A feature extractor for MIDI files. Proceedings of the International Computer Music Conference. 302–5.
Questions and Comments
DOWNLOAD FROM SOURCEFORGE
-top of page-