MuScriptor vs Omnizart

MuScriptor turns audio recordings into MIDI and sheet music, with downloadable MIDI files, sheet-music PDFs, and guitar-tab PDFs from its hosted service. The web demo accepts MP3, WAV, FLAC, OGG, M4A, AIFF, and Opus files. It can transcribe voice, piano, guitars, bass, strings, brass, woodwinds, drums, and other instruments, and users can specify which instruments they expect to improve transcription accuracy. Guitar and bass parts can include tablature alongside standard notation, but the system does not produce chord symbols. MusicXML output can be edited in MuseScore, Sibelius, or Guitar Pro. Users can self-host the web interface, and transcription is also available through a CLI and Python API. Local use requires Hugging Face authentication and acceptance of the model's CC BY-NC 4.0 license. Hosted audio is processed on the server and not retained afterwards; local use keeps audio on the user's machine. The service says it performs better on acoustic and guitar music than electronic music, and that results are not perfect. The hosted demo is free; the code is MIT-licensed, while model weights carry the non-commercial license.

Omnizart is a free Python library for transcribing polyphonic music into pitched instruments, vocal melody, chords, drum events, and beats. Pre-trained checkpoints allow transcription without first training a model. The project also documents dataset downloads, feature preprocessing, model training, transcription, and sonification, with MIDI, CSV, and text output formats. It lists separate applications for music, drums, vocals, vocal contours, chords, and beats, along with chord detection, MIDI export, and stem separation. Installation options include pip, Docker, and Conda, and a command-line interface is provided. Documentation links to a Colab notebook and a Replicate web demo. Omnizart supports Linux, Windows, macOS, and self-hosted setups, but is incompatible with ARM-based macOS because of underlying dependencies. The package requires Python 3.8 or later and below 3.15. Drum checkpoints can be used for transcription, though the drum model has an unresolved issue that prevents loss convergence when training from scratch. It is MIT-licensed and offers GitHub Issues for installation, compatibility, and usage problems.