Automatic Melody Harmonization with Triad Chords: A Comparative Study
This is an incremental study that provides a comparative analysis of existing methods for automatic melody harmonization, which is a domain-specific problem for music generation and computational creativity.
The paper tackled the problem of automatic melody harmonization by comparing several canonical methods, including template matching, hidden Markov models, genetic algorithms, and deep learning models, and reported results from an objective evaluation with six metrics and a subjective study involving 202 participants on a new dataset of 9,226 melody/chord pairs.
Several prior works have proposed various methods for the task of automatic melody harmonization, in which a model aims to generate a sequence of chords to serve as the harmonic accompaniment of a given multiple-bar melody sequence. In this paper, we present a comparative study evaluating and comparing the performance of a set of canonical approaches to this task, including a template matching based model, a hidden Markov based model, a genetic algorithm based model, and two deep learning based models. The evaluation is conducted on a dataset of 9,226 melody/chord pairs we newly collect for this study, considering up to 48 triad chords, using a standardized training/test split. We report the result of an objective evaluation using six different metrics and a subjective study with 202 participants.