Incipit display


See the Verovio Plaine & Easie Input page for some example incipits rendered by Verovio directly in the web browser, and the MEI Viewer for the rendering of a full score encoded in MEI.
Incipit searching
Incipit searching in the discovery interface of Muscat is built on the Themefinder search engine developed by the Center for Computer Assisted Research in the Humanities (CCARH) at Stanford University and the Cognitive and Systematic Musicology Laboratory at Ohio State University.
Themefinder searches look for melodies that match a pattern using so-called regular expression. One key idea behind Themefinder is that a melody can be represented in different ways, allowing for various levels of search fuzziness to be offered. In the Muscat implementation of Themefinder, the levels of fuzziness are:
- Exact Pitch: exact match of the diatonic pitch names
- Interval: match of 12-tone interval series
- Refined Pitch Contour: match of directional contours of steps (one or two semitones) and leaps (bigger than minor third)
- Gross Pitch Contour: match of directional contours regardless of the interval
The implementation of Themefinder in Muscat is directly embedded in the Solr index, making the incipit search fully integrated with all the metadata search options available in the advanced search of the discovery interface. The query interface includes a music keyboard data input with instant music notation rendering with Verovio.
The priority in Muscat was for Themefinder to be easy to understand and to use. For this reason, the query interface does not require any specific syntax to be known for using it, as the case in the original version of Themefinder. Instead, the user chooses the level of fuzziness from a drop-down list.

Incipit melodic similarity
The incipit search approach embedded in the advanced search of Muscat is very useful and brings valuable results, in particular thanks to the various fuzziness levels it offers and because of its integration with metadata search. The original version of Themefinder offers even more fuzziness features, for example with the use of wildcards. However, these advanced options require the underlying syntax to be known by the user. More importantly, the approach of Themefinder reaches limits when melodic sequences that are similar present differences that makes them unlikely to be matched with a reasonably simple Themefinder query, as illustrated by the incipits below.
In the field of Music Information Retrieval (MIR) there has been continuous research for developing algorithms that can cope with these types of differences and tools that are more “musically” informed. The next version of Muscat (5.0) will integrate the Monochord tool developed at the University of Utrecht under the supervision of Prof. Franz Wiering
The tool acts as an incipit similarity ranking with the results returned starting from the most similar incipit. Various parameters allow for the adjustment of the similarity focus (e.g. pitch or duration) and the balancing of them.
Click on the blue search icons to see a live demo of the Muscat new experimental incipit search! In each case you will see in the result list similar incipits that would not be matched with a simple Themefinder query or a with an ordinary search algorithm.
![]() |
RISM incipit 190016489-1.1.1 |
|
![]() |
RISM incipit 452509339-1.1.1 |
|
![]() |
RISM incipit 454013562-1.1.1 |
|
![]() |
RISM incipit 603001069-1.1.1 |
|
![]() |
RISM incipit 702006855-1.1.1 |
|