Audiveris/audiveris

What it solves

Audiveris solves the problem of transcribing printed or handwritten music score images into symbolic digital formats. This process, known as Optical Music Recognition (OMR), allows users to take a static image of music and convert it into a format that can be played back, edited, searched, or republished.

How it works

The software integrates an OMR engine with a manual editor. The engine uses a variety of techniques to recognize different musical elements: neural networks for fixed-size shapes, template matching for note heads, external OCR for text, and morphological closing for beams. Because 100% accuracy is rarely possible, the integrated editor allows users to adjust engine settings and manually correct any remaining transcription errors.

Who it’s for

It is designed for musicians, archivists, and researchers who need to digitize large volumes of musical scores (up to hundreds of pages) from real-world sources like the IMSLP library.

Highlights

• Supports large-scale scores with hundreds of pages.
• Combines automated neural network recognition with a manual correction interface.
• Exports data to the MusicXML 4.0 format.
• Cross-platform availability for Windows, Linux, and macOS.