concordia-harmony 1.0.0

No project description provided

Concordia: Music Generation Using An SMT-based Harmony Solver

Overview

Suppose you are composing some music. You have a beautiful melody, but you don’t know what chord progression to use. We propose Concordia, an SMT-Based harmony-solving framework that helps composers and songwriters brainstorm musical ideas and test them out right away. Given some constraints ahead of time (e.g. baseline, melody, which chord to use), Concordia will generate a harmony that satisfies the constraints. In the following sections, we will discuss the implementation of the harmony solver in more detail. This project made the following contributions:

• Describe how musical elements can be expressed as variables that an SMT-solver can understand.
• Provide an extensible framework for constructing constraints around musical constructs.
• Provide An open-source and easy-to-install implementation of the harmony solver, available on here on Github and as a package on PyPI. An quick demo can be found in example.ipynb.

Supported Constraints

Constraint	Description
Note specification	Use specific notes at a certain point in the harmony.
Chord specification	Use a specific chord/chord type at a certain position in the progression.
Chord set	Only use chords specified to compose the harmony.
Voice range	Soprano cannot sing a really low note; some instruments can only play a small range of notes.
Voice leading	You might want to restrict the max interval of a jump in the melody.
Chord progression	Some chords typically do not proceed to some other chords.
Idiomatic uses	Some music events happen with others (e.g. cadences, voice exchange).

Architecture

Changes

To push an update to PyPI, first update the version number in pyproject.toml, then push to GitHub.