Interindividual variations in skeletal muscle metabolism make comparative analyses difficult. In this study, we have addressed the issue of capturing the variability of metabolic performance observed during muscle exercise in humans by using an original method of normalization.Metabolic changes induced by various kinds of exercise were investigated using 31P magnetic resonance spectroscopy (MRS) at 4.7 T in 65 normal subjects (23 women and 42 men) and 12 patients with biopsy-proven muscular disorders. Large variations in the extent of PCr breakdown and intracellular acidosis were recorded among subjects and exercise protocols. For all the data pooled, the amplitude of mechanical performance accounts for 50% of these variations. When scaled to the work output, variations of PCr consumption account for 65% of pH changes through a linear relationship. This linear relationship was substantially improved (90%) when both variables were scaled to the square of work output performed (P1 and P2). By capturing most of the initial interindividual variability (90%), P1 vs. P2 relationship represents an ideal standardization procedure, independent of any anthropometric measurements. This relationship also discloses a significant link between the extent of PCr breakdown and intracellular acidosis regardless of exercise protocol. Moreover, changes in the slope of the P1 vs. P2 regression curve, as measured in old subjects and in selected patients, directly reflect alterations of energy production in muscle.