和聲複雜度在自動伴奏系統之應用與研究

Abstract

本研究為探討和聲複雜度與其應用至自動伴奏系統，我們利用實驗來驗證量化後的複雜度是否符合一般大眾聆聽音樂時所認知的和聲複雜度，利用量化後的複雜度去控制、調整歌曲伴奏的和聲進行之縱向織度以及橫向密度，使自動伴奏系統可以與旋律有更好的契合度。

此外，我們提出DCPA（Dual Chord Progressions Algorithm）自動伴奏系統，將自動伴奏系統分為調性分析、和聲分析、和聲選擇與音型伴奏系統四大部分，並以通用性為考量，設立基本功能和聲進行規則，讓樂曲的基本和聲進行有所規範。在音型伴奏系統中，使用者可以新增各式不同的音型伴奏，讓自動伴奏的成品聽起來更為豐富。

以DCPA自動伴奏系統為基礎，在後續研究中可以將和聲選擇和音型伴奏系統延伸，新增和聲進行規則和音型，以因應各種不同曲風的樂曲。

This study aims to explore the complexity of chord progressions and its application to the automatic accompaniment system. Experiment was conducted to verify the harmony complexity by general public. To make a better accompaniment with melody, we use harmony complexity to control music accompaniment texture. In addition, we propose DCPA (Dual Chord Progressions Algorithm) automatic accompaniment system, which is divided into four parts: key finding, chord analysis, chord select algorithm, and pattern accompaniment system. To improve the versatility of system, we establish the rule of chord function to make the chord progressions comply with the rules. In pattern accompaniment system, users can add a variety of different patterns accompaniment, so that the automatic accompaniment sounds richer. Researchers can make more chord-select rules and patterns to coordinate specific music style in the future work.