William Ziegler

Music information retrieval (MIR) is the interdisciplinary science of retrieving information relating to music. Music information retrieval systems help manage collections of music material and provide access to music collections. Although the first published work on music retrieval appeared a long time ago, only until recently has a significant amount of research been done on music retrieval (Chowdhury, 1998)
Music information primarily consists of seven facets. These facets include pitch, tempo, harmony, timbre, editing, text, and bibliography (Chowdhury, 1998)). Traditionally, most music information retrieval has involved text-based metadata tags to describe identifying features. These identifying features generally included the title of the song, the composer, and the performer. However, in terms of music retrieval, these text-based metadata did not accommodate to the needs of all information seekers. Searchers may not know the metadata info mentioned. For instance, information seekers often know the melody or rhythm of a song, but they do not know the title or performer of the song. Recently, query by humming (QbH) retrieval software has been developed in an attempt to solve these types of problems relating to music information retrieval.
Query by humming (QbH) is a music retrieval system that works differently than the original classification and retrieval systems that focus primarily on the title, artist, composer, and genre. Instead, QbH primarily applies to music with a distinct single theme or melody. The system involves taking a user-hummed melody, which is the input query, and it compares the hummed melody to an existing database. The system then returns a ranked list of music closest to the input query.
Musipedia is an example of a QbH system, but Musipedia uses a variety of input methods. Not only does it use input methods such as humming or whistling, but it also involves tapping a rhythm on the keyboard , using the computer keyboard as a piano keyboard to search melodies, and it includes a contour search, using Parsons Code to encode the music pieces.
Rainer Typke started Musipedia in 1997, and he has been developing the program ever since. He first called the music search engine Melodyhound and Tuneserver. But after adding the Wikipedia-like collaboration features, such as users editing and deleting existing entries, he decided to change the name to Musipedia. As the name suggests, the site is inspired by Wikipedia, and the site refers to itself as the “Open Music Encyclopedia.” It is a searchable, editable, and expandable collection of tunes, melodies, and musical themes. Also, since 2006, the Musipedia search engine can also be used for searching the entire internet for MIDI files. Musipedia locates the MIDI files that go into its search index by using the Alexa Web Search service, which has been available since December 2005.
Since Musipedia is a collaborative music encyclopedia, any user can modify the collection of melodies and enter MIDI files, bitmaps with sheet music, lyrics or some text about the piece, or the melodic contours as Parsons Code. Musipedia has different language options, which is can be beneficial to some users that do not speak English. The different languages that a user can choose for the site are French, German, English, and a Chinese language.
Some MIR systems only check for exact matches or cases where the search string is a substring of database entries. For such tasks, standard string searching algorithms are used. For example, Themefinder searches the database for entries matching regular expressions. In this case, there is still no notion of distance, but different strings can match the same regular expression. One of the most notable challenges regarding MIR is the fact that information seekers can have poor skills in reproducing music for their queries. This is why approximate matching can be beneficial in MIR, and with approximate matching it can be useful to compute an editing distance. Musipedia is an example of a system that computes an editing distance (Typke et al, 2005).
For example, the melodic contour search uses an editing distance, and because of this, the search engine finds not only entries with exactly the contour that is entered as a query, but also the most similar ones among the contours that are not identical. Similarity is measured by determining the editing steps (inserting, deleting, or replacing a character) that are needed for converting the query contour into that of the search result. Since only the melodic contour is relevant, one can find melodies even if the key, rhythm, or the exact intervals are unknown.
The pitch and onset time-based search takes more properties of the melody into account. This search method, which is the default method, is still transposition-invariant and tempo-invariant, but it takes rhythm and intervals into account. The melody can be entered in various ways, but the primary way is by clicking on a virtual keyboard on the screen. The search engine then segments the query, converts each segment into a set of points in the two-dimensional space of onset time and pitch, and it then compares each point set to pre-computed point sets representing segments of melodies from the database. As with the contour search, little alterations of the query will lead to correspondingly small changes in the results, which helps when there is a user error when the user submits a query (Wikipedia, 2011).
The query by tapping method only takes the rhythm into account, and the user taps in the rhythm with any key on the keyboard. The tapping method uses the same algorithm as the pitch method, but assumes all pitches to be the same.
For more technical information as to how the melody search tool works, a user can view citations of various articles written by Rainer Typke. This can be found by clicking on a link at the bottom of Musipedia’s “About” page. After clicking on the link, a user can view published works by Typke from 2001 to 2011. Most of the content of the articles are notably technical and can be difficult to understand for those without considerable computer programming knowledge.
Musipedia does not always seem to be accurate, and the collection of music that Musipedia pulls from can often seem limited. However, since people usually do not memorize everything about a song the first time they hear it, and since information such as title, composer and performer are often learned at a much later stage of a person’s relationship with a song, Musipedia and other QbH systems seem to be information retrieval tools that are leading the way to more practical, and more encompassing music retrieval systems.

Chowdhury, G. (2010). Introduction to modern information retrieval. (3 ed., pp. 348-349). London, England: Facet Publishing.
Musipedia. Retrieved from
Typke, R., F. Weiring, & R. Veltkamp (2005). Survey of music information retrieval systems. ISMIR, 153-160.
Wikipedia: article on Musipedia. (2011) Retrieved from