Defining Poor-Pitch Singing: A Problem of Measurement and Sensitivity

Important individual differences characterize singing in adult nonmusicians. In spite of the fact that the majority can carry a tune, some occasional singers are particularly inaccurate or imprecise in producing or imitating pitch (poor-pitch singers). Poor-pitch singing can be defined via acoustical analyses of vocal performances. In spite of the objective nature of this method, however, to date there is not a standard strategy for identifying and describing poor-pitch singers. Different tasks (e.g., singing from memory vs. imitation), cut-offs (50 cents, 100 cents, vs. variable criteria), and metrics (e.g., accuracy vs. precision) are typically used for assessing singing proficiency. Here the pros and cons of different methods and measurements are discussed. The boundary between poor-pitch singing and good singing depends on these factors, which should be carefully taken into account when assessing singing abilities. An approach based on multiple tasks sharing a common set of measures of singing accuracy and precision, with a relative cut-off (i.e., 2 SD from the average of a normative group) is favored for identifying poor-pitch singers.