The Optical Neume Recognition Project
A tool to investigate early staff-less music notation

Ven Venite et videte locum ubi positus erat dominus alleluia alleluia (cao5352)
St. Gallen, Stiftsbibliothek, Cod. Sang. 390, p. 32 (www.cesg.unifr.ch)

Related Projects

Several previous attempts at combining OCR and neumes should be mentioned here. First, the beginnings of an e-library for chant transcriptions have been developed by Barton, Caldwell and Jeavons as part of the NEUMES project http://www.scribeserver.com/NEUMES/. The goal of this project is to design a software infrastructure (using an XML Schema) for the digital transcription of chant manuscripts, so that they are readable across many types of applications programs and on the WebLouis W.G. Barton, John A. Caldwell, Peter G. Jeavons, E-Library of Medieval Chant Manuscript Transcriptions in JCDL (Denver, Colorado, 2005).. NEUMES is interested in a standardizing the neume forms and creating new documents from them, and is of the view that since early Gregorian notation is not standardized, OCR is of limited benefit. Our project focuses precisely on that which NEUMES considers a flaw, that is, we attempt to recognize each shape in whatever way it has been drawn, to get a clearer understanding of the breadth of variance allowed within a particular type of neume notation. The current status of NEUMES is unknown; the last copyright on the webpage is 2007.

Using a document recognition system closer to what we propose here, psaltic Byzantine chant notation has recently been the subject of another study by Dalitz, Michalakis and Pranzas Christoph Dalitz, Georgios K. Michalakis and Christine Pranzas, Optical Recognition of Psaltic Byzantine Chant Notation, International Journal of Document Analysis and Recognition vol. 11, no. 3 (Dec. 2008): 143–158.. The particular type of notation used here was standardized in the early 19th century, lending itself to the process of optical recognition based on the Gamera framework for document image analysis.

One other experiment, undertaken twenty years ago, involved the application of optics and image processing on late medieval notation. In an article entitled "The Optical Scanning of Medieval Music", a computer scientist and a musicologist at the University of Ottawa (McGee and Merkley) proposed eliminating the staff lines in images of square notation which would increase the computer’s ability to recognize graphic patterns William McGee and Paul Merkley, The Optical Scanning of Medieval Music in Computers and the Humanities 2 5: 47-53, 1991.. They found that the graphic result of stripping away the staff lines is "to transcribe the square neume into something very much like its cheironomic counterpart" Ibid, 51. and, later, "we can here see the full … interconnection between the early symbols and the later system of heighted pitch notation … [when we] peel back the historical layers of the notation … in the process of optical scanning." Ibid, 51. Our project, of course, begins with these earliest forms of neumes (which, as mentioned above, we do not attempt to standardize graphically) — the very neumes at which the researchers in late medieval notation eventually arrive!