Calliope: A Web MIDI Environment for Computer-Assisted Music Composition
Calliope is a web-based interactive environment for computer-assisted composition offering access to a version of the Multitrack Music Machine (MMM) model for symbolic music generation. The user can generate or re-generate symbolic music using a seed MIDI file via a practical easy-to-use graphical user interface (GUI). The system can interface with your favorite DAW (Digital Audio Workstation) such as Reaper via MIDI streaming.
Non-Disclosure Agreement (NDA): This is a beta-testing project. By accessing and using the Calliope software and its documentation, you agree not to distribute or communicate about it to anyone but those directly involved with this project. We collect in-app user data including account information solely, for communication and research purposes. We take your privacy seriously and we will anonymize. Additionally, journalists and media are not allowed to communicate about the project at this time. Please, contact us directly for further inquiries.
Demo
This is a short demo video for press. The MIDI piece used is Mozart’s K333 Piano Sonata 3rd Movement (edited to get the first 32 bars only).
Artwork
Computer-Assisted Composition in Calliope
Calliope is used to generate new MIDI content from existing MIDI files or to generate from scratch. This is achieved by selecting bars to be re-generated and a destination instrument which will play the generated content (e.g. trumpet, drums, violin or piano). This influences the model generation because composing for a drums section is a different task than composing for a string section. Calliope offers a unique creative workflow where you can generate variations in batch. This allows to choose from many generations. Calliope is aimed at both amateur and professional composers/users. It is an outlet for exploring new creatively generated ideas for your next music project. It is also useful for experiencing alternative workflows to traditional music composition and production, and for exploring AI possibilities in digital music composition.
Computer-assisted composition is a field that studies the partial or complete automation of music composition tasks. Calliope supports many of such tasks including in-painting multi-track pattern generation, multi-track complete generation, rhythm generation, harmonization, chord progression generation, melody generation and interpolation. The research field employs artificial intelligence (AI) techniques such as machine learning (ML), deep learning (DL), evolutionary methods and/or rule-based systems to model musical information and intelligence.
Features
Available online 24/7 with multi-user bandwidth and user authentication support
MIDI viewer and editor with bar selection for a generation
MIDI playback (soundfonts) and streaming (e.g. to a DAW such as Ableton)
Multi-track music generation with global and track-specific control parameters
Batch generation of several MIDI variations
Automated ranking generated MIDI files against an original collection
Exporting and streaming MIDI files
Releases
Google Chrome ONLY!
Calliope v0.11 – Feb 12th 2024 (User Guide)
Calliope v0.10 – Feb 8th 2023 (User Guide)
Calliope v0.9.5 (dark themed) - June 6th 2022 (User Guide)
Calliope v0.9.5b1 - May 1st 2022
Apollo v0.9.4a1 (Calliope) - Jan 11th 2022
Apollo v0.5.0 pre-alpha - Sep 17th 2021
Apollo v0.1 - Jun 17th 2019
A detailed history of release activities can be found here. Calliope is an evolution born out of the (now archived) Apollo project.
MIDI resources
The following resources contain MIDI files that can be used in Calliope:
Papers and Posters
Bougueng R. T. and Ens J. and Pasquier P. "Calliope: An Online Generative Music System for Symbolic Multi-Track Composition". Peer-reviewed Demo at International Conference on Computational Creativity (ICCC) 2022.
Bougueng R. T. and Ens J. and Pasquier P. "Calliope: A Co-Creative Interface for Multi-Track Music Generation". Peer-reviewed Demo Paper at ACM Creativity and Cognition (C&C) 2022.
Bougueng R. T. and Ens J. and Pasquier P. "Calliope: Generating Symbolic Multi-Track Music on the Web". Demo Paper at Sound and Music Computing (SMC) 2022.
Bougueng R. T. and Ens J. and Pasquier P. “Apollo: An Interactive Web Environment for Generating Symbolic Musical Phrases using Corpus-based Style Imitation”. In Proceedings of the 7th International Workshop on Musical Metacreation (MUME 2019). Vol. 7. June 2019.
Ens J., Pasquier P. “MMM: Exploring Conditional Multi-track Music Generation with the Transformer”. arXiv preprint arXiv:2008.06048. 2020 Aug 13.
Ens J. and Pasquier P. “Quantifying Musical Style: Ranking Symbolic Music based on Similarity to a Style”. In Proceedings of the International Symposium on Music Information Retrieval. Vol. 20. 2019, 870-877. (Oral Pres.)