论文标题
超越和弦词汇：在和弦估计中利用音调关系度量
Beyond chord vocabularies: Exploiting pitch-relationships in a chord estimation metric
论文作者
论文摘要
和弦估计指标将和弦标签视为彼此独立的。这无法以有意义的方式代表和弦之间的音调关系，从而进行了评估，这些评估必须使复杂的和弦词汇表构成妥协，并且通常需要进行耗时的定性分析，以确定和弦估计算法如何表现详细信息。本文提出了和弦估计的准确度指标，该指标将估计和弦的音调含量与地面真理进行了比较，该真相既捕获了估计的正确说明又是插入估算值的其他音符。这不是独立的评估协议，而是可以作为权重整合到现有评估方法中的度量。
Chord estimation metrics treat chord labels as independent of one another. This fails to represent the pitch relationships between the chords in a meaningful way, resulting in evaluations that must make compromises with complex chord vocabularies and that often require time-consuming qualitative analyses to determine details about how a chord estimation algorithm performs. This paper presents an accuracy metric for chord estimation that compares the pitch content of the estimated chords against the ground truth that captures both the correct notes that are estimated and additional notes that are inserted into the estimate. This is not a stand-alone evaluation protocol but rather a metric that can be integrated as a weighting into existing evaluation approaches.
