What if music could wait for its listener?
Not for years. Not for centuries. But for a whole millennium.
And one day, sound again.
This is precisely what resulted from the European research project REPERTORIUM, which officially launched on January 1, 2023, under the Horizon Europe program.
The project's name stands for Researching and Encouraging the Promulgation of European Repertory through Technologies Operating on Records Interrelated Utilising Machines — the research and preservation of European musical heritage using artificial intelligence and digital technologies.
For three years, an international team of researchers from Spain, Italy, France, Germany, Finland, the Netherlands, Lithuania, and the UK brought together musicologists, historians, and specialists in artificial intelligence and digital technologies.
Their goal was unusual.
Not to create new music.
But to help the modern world hear again that which had almost disappeared from cultural memory.
The foundation of the project was the unique collection of the Atelier de Paléographie Musicale — the musical palaeography workshop of the Solesmes Abbey in France.
For over a hundred years, Benedictine monks collected and photographed medieval musical manuscripts from libraries across Europe. Over time, this collection grew into one of the most important archives of Gregorian chant — a musical tradition that had a profound impact on the development of European culture.
However, studying such an archive manually is practically impossible.
This is precisely where modern technologies came to the rescue.
Using methods of digital image processing, optical character recognition for ancient notation, and artificial intelligence algorithms, researchers digitized tens of thousands of manuscript images and created an open digital platform, Medieval Music Manuscripts Online (MMMO). Thanks to this, thousands of medieval sources became accessible to researchers worldwide, and many forgotten chants were systematized and identified for the first time.
But the most significant outcome of the project was not the technologies themselves.
In January 2025, Gregorian chants, which researchers estimate may have been silent for over a thousand years, were performed for the first time in the modern era at the Cathedral of Salamanca.
They were performed by the ensemble Schola Antiqua, led by Juan Carlos Asensio Palacios, one of the leading specialists in medieval music.
Later, on November 21, 2025, these works were released on a double album titled Beatæ Mariæ: Canto llano en la Catedral de Salamanca.
Thus, music that had existed solely in ancient manuscripts for centuries once again became part of the living cultural space.
And although the project itself concluded at the end of 2025, its story continues.
Today, all digitized materials and research results are available on the open platform Medieval Music Manuscripts Online (MMMO), supported by DIAMM (Digital Image Archive of Medieval Music) at Oxford University. This allows musicians, researchers, and anyone interested in medieval music to freely work with unique historical sources.
This research also offers a new perspective on the role of artificial intelligence.
Most often, we speak of it as a tool for creating new things.
But in this case, it served a different role.
Not as a composer. Nor as an author. But as a restorer of cultural memory.
What did this event add to the world's soundscape?
Behind every new algorithm, we often see only the technology. But sometimes, technology becomes a bridge. Not between humans and machines. But between the past and the present.
The REPERTORIUM project demonstrated that artificial intelligence is capable of not only creating new things.
It can help humanity hear again what was already created centuries ago.
And perhaps, the most astonishing discovery lies not in artificial intelligence at all.
But in the music itself. It outlived its composers. It outlived epochs. It outlived centuries.
And all this time, it seemed to be waiting for the moment when it would be heard again.
If music is indeed capable of waiting for its listener for hundreds of years…
What other voices from the past are still waiting for us to learn to hear them?
