Over the years, a growing number of researchers have been developing models that would automatically generate assessments of music performances. Yet the number and usage of automatic singing evaluation systems is still rather rudimentary, addressing, for the most part, a limited amount of performance features and lacking verification. This study reports on a newly designed automatic singing assessment tool based on a score-based model and its validation. Short music segments (N = 2640) were gathered via recordings made by music education students (N = 55) of a specially inscribed vocal music excerpt. Recorded data evaluation was generated by a specially devised automatic tool as well as by three human experts, addressing pitch intonation (examined for its overall display, single note accuracy and interval manifestation), dynamics transmission and vocal resonation quality. Findings indicated a higher rating given by the experts in regard to pitch intonation and vocal resonation. However, a similitude was found for the dynamics transmission scoring, and a correlation was found for pitch intonation and the dynamics transmission scoring level: in both performance parameters, the higher the experts’ gradings were, the higher the gradings provided by the automatic tool. Results attest to the automatic tools’ qualification as an aid for human judgement of singing proficiency. The tool could assist investigations in various musical domains, such as music pedagogy, music performance or music perception research.