Automated Tessituragram Analysis Software of Musical Instrument Digital Interface (MIDI) Files to Quantitatively Assess Musical Demands.

Abstract

Objectives: Quantitatively assessing musical demands primarily relies on time-consuming manual data entry methods. This reliance limits the development of the larger datasets needed to assess generalizable trends in musical compositions. To remedy this labor-intensive manual data entry issue, the purpose of this study was to: DESIGN: A Python application was developed to parse and convert single-track MIDI files into a list of pitch frequencies and silences. The software then uses this list to compute musical difficulty metrics based on previously established tessituragram analysis methods. After parsing a .midi or .mid file, the software outputs the total performance time, time dose, rest time, cycle dose, total range, interquartile tessitura range, and the percent of time spent in generalized musical passaggi as a user-friendly graphic with numeric depictions or as a numeric table. The software was tested using MIDI files generated from the transcribed melodic lines of 150 manually assessed operatic arias.

Results: Comparing software outputs against the manual analysis of the 150 arias reveals a 98.7% aggregate match, indicating that the software is accurate. Differences between software and manual measurements are likely attributable to input errors during human data entry and the software's ability to parse the data more finely.

Conclusions: The software developed in this research quantitatively analyzes single-track MIDI files and outputs various metrics of musical demands. Outputs from the software align with the results of previously published research using manual data entry methods. Future scholarship will utilize this software to generate larger datasets and assess generalizable trends in musical compositions.