The influence of music on physiological and affective exercise responses was studied in 8 trained and 8 untrained runners under three music conditions ("no", "sedative", and "fast") during low, moderate, and high intensity exercise. Repeated measures ANOVA revealed increased respiratory frequency (FR) during fast music as compared to the no music and sedative music conditions (p < 0.01). Plasma cortisol levels did not differ at baseline across the music conditions; however, following high intensity exercise, higher cortisol levels were associated with fast music as compared to no music and sedative music (music x intensity interaction, p < 0.01). Affective measures during exercise (FEELING scale) showed no overall training group differences; however, there was a music x group x intensity interaction (p < 0.05) in which untrained subjects reported more positive affect compared to trained subjects while listening to fast music during low and high intensity exercise. Data collected at voluntary exhaustion revealed significantly more positive affect and higher skin temperature (p values < 0.01) in untrained compared to trained subjects. Collectively, these results suggest listening to fast, upbeat music during exercise may be beneficial for untrained runners but counterproductive for trained runners.