Music Information Processing Using the Humdrum Toolkit: Concepts, Examples, and Lessons

Humdrum is a general-purpose software system intended to assist users in a variety of music-related applications. This article provides an introductory tour of Humdrum emphasizing three goals: to identify the basic elements, operations, and organization of Humdrum; to illustrate the range of applications through a series of tutorial examples; and to identify a number of lessons from Humdrum that might benefit future music software designers.

Humdrum's capabilities are quite broad, so it is difficult to describe concisely what it can do. Since its inception, the principal users of Humdrum have been systematic musicologists, theorists, ethnomusicologists, music librarians, historians, music cognition researchers, and composers. Five features have accounted for this broad interest: the ability of users to concoct or tailor unique representations that pertain to the user's specific interests; a flexible set of analytic and processing tools that can be applied to both established and user-defined representations; a coherent and extensible system for representing reference-related metadata; ease of connectivity to other existing software; and availability of a large volume of high-quality encoded materials.

In Humdrum, new musical representations are easily created. The opportunity to craft representations for specific tasks has led to the development of innumerable representations. For example, Humdrum users have concocted representations for Bugandan xylophone music, Cajun button accordion tablatures, square notation, Persian Ney music, Benesh dance notation, running acoustic spectra, and track index markers for compact discs. For many projects, it is common to generate intermediate or "throw-away" representations that are used only for a single task. For example, in perceptual research, collected data (such as listener responses) are commonly encoded in the same document that contains the stimulus materials.

In addition to specific representation schemes for user "data," Humdrum allows users to define their own types of reference-related information. Conventional reference information includes an artist's name, title, date of performance, copyright status, etc. However, Humdrum allows users both to extend and omit such metadata in ways that maintain compatibility without compelling users to encode task-irrelevant materials. For example, a ballet scholar might define reference information identifying the transcriber of a Laban dance notation while remaining compatible with conventional cataloguing systems. At the same time, users can still process materials with incomplete or absent reference information.

Although Humdrum does not provide its own sound generation or graphic notation capabilities, a number of translators exist that connect Humdrum to existing applications. These include translations to MIDI, Csound (Vercoe 1986), the Finale Enigma format, the Score notation system (Smith 1972; Kornstädt 1996), and the Mup music text formatter (available from www.arkkra.com). In addition, materials can be translated to the Humdrum format from MIDI, the Finale Enigma format, MuseData (Hewlett 1997), DARMS (Erickson 1976; Hall 1997), and the RISM Plaine and Easie representation (Howard 1997).

All Humdrum file formats are ASCII text. This facilitates viewing and editing data, increases cross-platform portability, and eases the writing of task-specific software. The plain text format also facilitates reformatting data for special-purpose uses such as graphics programs or statistical packages.

Musical materials encoded in the Humdrum format span five centuries and six continents. Encoded materials include complete musical works as well as collections of themes and incipits. As of 2001, over 40,000 musical works had been encoded in various levels of detail. Some encodings are simple monophonic melodies; others include full orchestral scores including such details as stem-direction and beaming. In addition to the MIDI format, the Humdrum "kern" representation is one of the principal distribution formats for the high-quality electronic editions produced by the Center for Computer Assisted Research in the Humanities (www.ccarh.org). An extensive collection of non-Western musics has been encoded; an incomplete alphabetic sample includes Aleutian, Brazilian, Chinese, Dutch, Ethiopian, Fijian, German, Haitian, Indonesian, Japanese, Khmer, Lakota, Mandan, Namibian, Ojibway, Peruvian, Quebecois, Russian, Scottish, Thai, Usarafus, Venda, Xhosa, Yoruba, and Zulu musics. (Huron 1992). Regarding Humdrum's analytic capability, a selection of tools will be discussed next.

Humdrum Syntax

Humdrum data are organized in two-dimensional tables like a spreadsheet. Columns of data...