The RISM Database of Musical Manuscripts in US Libraries

RISM (Répertoire Internationale des Sources Musicales, or International Repertory of Musical Sources), series AII, is a database of music manuscripts. It is international in scope. The site, which is limited to American holdings, is operated by the Center for Computer Assisted Research in the Humanities at Stanford University in cooperation with the US RISM Committee and the US RISM office at Harvard University. The emphasis has been on holdings of music from the seventeenth and eighteenth centuries in the libraries of universities, conservatories, museum, and private collections within the United States. Eventual coverage of nineteenth century materials is intended.

1. The Central Purpose of RISM

RISM AII series database is different from most bibliographical databases in than it combines textual and musical content. Many pieces of music were copied over and over but with small differences from time to time. The RISM musical content is valuable for many reasons: By culling works with similar musical properties, new concordances and attributions, as well as new disputes over authorship, sometimes emerge.

2. The US RISM/Themefinder website

This site ( features the data for musical manuscripts from the United States. At present there are 55,491 listings. (Entries lacking musical incipits are excluded here.) More than 100 fields track diverse details of each manuscript. The content is textual in all but one of them.

2a. Text search

The data can be searched by any of the several dozen text fields found in the metadata. The most commonly sought ones (composer, title, genre, keyword (all other metadata fields), and library have independent search boxes. Full entries of metadata may be generated by clicking the i button under the musical incipit. The field numbers at the left observe bibliographical norms adopted by the US RISM committee. Field entries may include information provided by cataloguers and encoders. When present, it is placed in square brackets ( [ ] ). These may interfere with text searches. For example, the name of Bachs son Carl Philipp Emanuel (commonly cited as C.P.E. Bach), is given in several instances as Bach, [Carl Philipp] Emanuel.

2b. Music search

The Themefinder lookup is optimized for both literal and fuzzy queries of musical data. The database does not include complete works. It stores only the beginnings of musical lines for vocal and instrumental parts. The musical search boxes allow for fuzzy searching of musical incipits. Fuzzy searching is useful in finding concordances in which key, details of meter or rhythm, and placement of rests may prevent identification in literal searches. Five levels of detail are supported. In multi-movement works, each movement has a separate entry. Work movements often become separated in manuscript transmission, so only a single movement from a multi-movement work may survive in a given manuscript. For further information, consult the Thematic catalogue field, which cites references works devoted to single composers, collections, or repertories.

2c. Results page

Icons under each musical incipit on the results page extend the uses of the material. The i button displays all the metadata for one record. MIDI files are terminated to conform in length to the musical incipit shown. The clipboard addition button, + allows users to save selected entries. The collection facility, C, shows all the other entries in a volume of diverse works described by a single RISM number.

2d. The Themefinder music clipboard

Entries on the clipboard may be rearranged by using these icons: Other icons on the clipboard have these meanings: View Humdrum data

2e. Copies of the manuscripts cited

Those wishing to obtain a photocopy of a specific manuscript should contact the holding library directly.

3. Technical information

3a. The Themefinder search engine

Music queries are run on data converted from Plaine & Easie format into the Humdrum format. The Humdrum view is shown as ASCII text, with the original RISM field codes and contents (prefaced by as comments beginning !!!RISM-) prefixed to the musical-data translation from Plaine & Easie Code into Humdrums **kern code.

For a brief introduction to Humdrum (which is used for musical-data query), see these sites:

A list of sources for fuller information on Humdrum can be found in

Summary encoding principles for both Plaine & Easie Code and Humdrum are given in Beyond MIDI: The Handbook of Musical Codes (The MIT Press, 1997). See

3b. RISM/Themefinder error reporting

The R button in the Information (i) view brings up a screen for reporting errors (which is currently inactive). Suggested changes (here captured in the Humdrum format) are sent automatically to the US RISM Office,

General feedback on the Themefinder interface can be given at the feedback link found on the Results page.

3c. Other RISM-Catalogued Manuscript Collections with Online Access

A composite collection (including the US data) is currently available via a CD-ROM which can be licensed from K. G. Saur (Germany). Details may be obtained from the RISM Central Office (Zentralredaktion), The US portion constitutes less than 10% of all the entries managed by the Central Office in Frankfurt (Germany).

Other national collections with individual websites include these:

These facilities permit literal search of the incipit code (which is called Plaine & Easie).

4. Notes and Credits

4a. RISM AII Database Content

The RISM encoding of titles and of musical incipits provides a facsimile of what is in the manuscript itself. This may include abbreviations, words in archaic spellings, and musical measures with missing or superfluous beats. The data is provided on an as is basis. Spelling variations and notational errors may occasionally prevent the identification of matches. Genre classification is based on textual data, not musical content.

4b. The Themefinder Musical-Search Engine

The Themefinder search engine for musical content has been under development since 1996 at the Center for Computer Assisted Research in the Humanities at Stanford University. Most of the development and maintenance has been carried out by Craig Stuart Sapp. David Huron initiated the idea. Andreas Kornstädt developed the original interface. A number of Stanford students have contributed to data encoding and assembly. Non-RISM portions of Themefinder include lookup facilities for several collections of folk music, classical music of the 17th through early 20th centuries, and Latin motets from early printed sources. For access to repertories other than the RISM manuscript collection, please go to and select a different repertory. Alternatively, you may search the entire collection by selecting ALL.

4c. The Center for Computer Assisted Research in the Humanities

Founded in 1984, the Center for Computer Assisted Research in the Humanities is engaged in the development of large databases of musical and textual materials for applications in research, teaching, and performance. Information on and links to its data holdings and materials on musical informatics can be found at

Last updated: 11/24/08