Digital projects / Canadian researchers collect data from Gregorian chants in the Convent of St. Gall, Switzerland

A new Canadian research project is collecting big data from medieval melodies chanted by monks more than 1,000 years ago. And it’s all searchable. But to what end?

Kate Helsen, an assistant musicology professor at Western University’s Don Wright Faculty of Music, is part of the Optical Neume Recognition Project and explains that this study is the most technologically advanced method of investigating what was previously a completely oral culture – a time and place, when and where people didn’t conceive of writing music down at all – and through greater understanding of these 11th century monks, researchers can now study how the human brain constructs, comprehends and reconstructs everything from language and literature to math and music.

The Optical Neume Recognition Project uses modified optical character recognition (OCR) technology to study medieval musical notation called neumes. This unique computing initiative identifies each neume on a digital scan of a page from a historic book of medieval chant, cataloguing each one and ranking them in order of graphic similarity while additionally storing data about how often a particular neume appears in specific combinations thereby creating a virtual ‘dictionary’ of neume signs.

Instead of poring over hundreds of pages of literally millions of neumes, researchers can now electronically search for information that has been systematically gathered by advanced computer software about each scanned image, using optical music recognition (OMR) software.

“Basically, we are mining these melodies for a better understanding of how the brain breaks down, thinks about and reconstructs melody year after year after year in a monastic context because that’s what was important to them. To sing the same prayer, the same way every year,” explains Helsen, who is currently attending the Music Encoding Conference in Montreal. “A lot of medieval scholars think that it’s not possible to retain all of that information. They’d say that we just haven’t found the books yet but I disagree. The medieval memory was fabulous for a lot of reasons and this is just another example.”

The research team is currently investigating the Gregorian chants of monks from the Convent of St. Gall in St. Gallen, Switzerland, which is considered a prime example of a great Carolingian monastery. It is estimated that it would take 85 hours to sing the entire prayer cycle of the St. Gall monks. But 1,000 years ago, this entire repertoire was memorized and repeated year after year by the monks and this continued for another 200 to 250 years.

“This memory work was done completely without notation. It was all in their heads. Today, we don’t have to have 85 hours of repertoire in our head because we have notation,” says Helsen. “And by developing a searchable database, not unlike Google Books, we are basically creating an electric monk. A device that knows all of the melodies. It’s as though a monk from 1,000 years ago walked into the room and started talking about music. It’s all there.”

The Optical Neume Recognition Project, currently funded by Social Sciences and Humanities Research Council (SSHRC), is part of a larger research collaboration known as Single Interface for Music Score Searching and Analysis (SIMSSA), which seeks to create a single clearinghouse for digital images of musical scores, both printed and handwritten, around the world.

Via Press release.

Advertisements