Visual Analysis of Voice in Crossover Singing

The current research on acoustic properties in singing voice analysis has mainly focused on individual song segments, and analysed them using simple data tables and basic charts. However, there has been limited exploration of comparing data from multiple sources, and visual analysis in crossover singing has been either too simplistic or too complex to provide a comprehensive view. This study aims to address this gap by incorporating song segments from different musical styles and utilising an innovative graph drawing method to generate interactive graphs for comprehensive musical data analysis. The findings provide additional support for existing statements in the field of musical data analysis, and demonstrate the effectiveness of the proposed graph method for analysing multiple song segments. At this stage, the study's findings confirm that the formant frequency
$F_{1}$
of singing across styles is less modified, but
$F_{2}-F_{5}$
varies in styles singing in English. Additionally, the formant frequency in Mandarin and Cantonese singing may be associated with pitch. The study also identifies that visualised graphs can produce similar results as current vocal research and are convenient for reading multiple data simultaneously. The methodology has the potential to be extended to other areas of musical visualisation to uncover insights from complex music datasets.