The evaluation of singing voice accuracy: a comparison between subjective and objective methods

Area of Science:

• Music Performance Science
• Acoustic Phonetics
• Vocal Pedagogy

Background:

• Vocal accuracy in singing is traditionally assessed through subjective human judgment.
• Objective acoustic analyses offer quantifiable metrics for vocal performance evaluation.
• A direct comparison between these two assessment methods has been lacking.

Purpose of the Study:

• To directly compare the reliability of acoustic analyses versus subjective judgments in assessing singing vocal accuracy.
• To identify specific acoustic parameters that correlate with expert perception of vocal accuracy.
• To validate the use of acoustic analysis as a complementary tool for vocal assessment.

Main Methods:

• Recorded 166 untrained singers performing 'Happy birthday'.
• Conducted acoustic analyses to measure pitch interval deviation, contour errors, and tonality modulations.
• Obtained global pitch accuracy ratings from 18 singing/music experts.

Main Results:

• A high correlation was found between acoustic measurements and expert subjective ratings.
• Acoustic analyses explained 81% of the variance in judges' scores.
• Subjective ratings were significantly influenced by pitch interval deviation and tonality modulations.

Conclusions:

• Objective acoustic measurements and subjective vocal assessments show significant congruence.
• Pitch interval deviation is a crucial acoustic criterion for vocal accuracy assessment.
• Tonality modulations are important for perceptive rating and should be included in acoustic analyses.