AnthemScore by Lunaverus

Using machine learning trained on musical audio, it detects notes, beats, and instruments with a level of accuracy that makes it a practical tool for real-world transcription projects rather than just a novelty.

The transcription process is designed to be interactive rather than a black-box one-click output: users can adjust a slider to intelligently add or remove notes based on likelihood scores, letting the software suggest candidates while giving musicians final editorial control over the resulting sheet music.

This collaborative approach between AI and human judgment produces more accurate results than fully automated transcription, particularly for complex polyphonic recordings where note separation is ambiguous.

Advanced editing capabilities distinguish AnthemScore's Professional and Studio editions from the entry-level Lite version: a spectrogram view shows the frequency content of the audio alongside the notation, a piano roll timeline display provides an intuitive visual representation of note durations, and a keyboard view connects the notation to a visual instrument layout.

These views allow musicians to cross-reference the audio signal with the generated notation to catch and correct transcription errors efficiently.

Instrument-specific formatting is handled automatically: users select the target instrument, and AnthemScore formats the sheet music in the appropriate key, range, and staff layout for that instrument.

Copy-paste of musical sections, downbeat tapping for tempo synchronization, and playback speed adjustment for detailed inspection complete a workflow that covers everything from initial transcription through to final notation refinement.