Automatic Singing Performance Evaluation for Untrained Singers

Abstract

SUMMARYIn this letter, we present an automatic approach of objective singing performance evaluation for untrained singers by relating acoustic measurements to perceptual ratings of singing voice quality. Several acoustic parameters and their combination features are investigated to find objective correspondences of the perceptual evaluation criteria. Experimental results show relative strong correlation between perceptual ratings and the combined features and the reliability of the proposed evaluation system is tested to be comparable to human judges.

Deep Learning Approaches in Topics of Singing Information Processing

“…Cao et al [33] summarized five perceptual evaluation criteria for assessing non-trained singers. Those perceptual parameters were: intonation accuracy, described as singing in tune, where suitable key transposition is allowed; rhythm consistency, described as singing with appropriate tempo speed, where slight tempo variation is allowed; timbre brightness, described as brilliance of tone, a sensation of brightness of spectrum; vocal clarity, described as vocal vibrations of a clear, well-produced tone; and overall performance, described as the overall evaluation integrating all perceptual parameters.…”

“…In another study, Tsai and Lee [32] proposed an automatic evaluation system for karaoke singing in which they computed the Euclidean distance of the note sequence of test singing voice from that of the intended reference song, time-aligned with each other using DTW, to compute the pitch accuracy rating. Other reference-dependent techniques [13], [33], [45] have also been proposed.…”