Although there is a considerable body of research on factors influencing musical emotion induction, relatively little is known about how the quality of performances may affect listeners' music-evoked emotion. To redress this gap, we generated music excerpts performed at three levels of proficiency (plain, semiprofessional, professional) and studied the effects of these variations on liking and on felt emotion via the Geneva Emotional Music Scale-a domain-specific instrument for the assessment of music-induced emotions. After pretesting the stimuli, in Study 1 (N = 207) we examined pairwise differences between each of these conditions, as well as interaction effects with musical expertise. In Study 2 (N = 143), each participant heard all three versions of a selection of music excerpts used in Study 1. Using multilevel modeling, in Study 2, we also examined interaction effects between performance and listener attributes that are known to modulate musical emotion induction, such as musical expertise, musical aptitude, personality factors, and current mood state. Across the studies, the results converge in showing that liking and intensity of emotion increased as a function of performance quality. In addition to performance main effects, we also found Performance × Listener Features interactions, such that the emotional effect of the performance condition was enhanced in participants with higher musical expertise and aptitude. By particularizing the role of performance features, the current study closes an important gap in understanding factors involved in musical emotion induction.