The effects of musical context and musical instrument characteristics on timbre discrimination in a concert hall were examined. Isolated tones and six-tone melodies were recorded in two different locations of a concert hall using five musical instruments. Each trial consisted of either a pair of tones, or a pair of melodies, which had been recorded simultaneously at the two locations on the same instrument. No significant differences were found between responses for isolated tones and responses for melodies. Musical instrument characteristics did prove to be significant, with listeners best able to discriminate flute tones, followed in order by clarinet, trumpet, viola, and marimba. In a follow-up analysis, five physical attributes of each instrument were tested for their contribution to the predictability of the discrimination scores. Variance with respect to the mean power level of the sustained portion of a tone was found to be the dominant contributor of the total variance of subjective responses, and this implicates differences in the early reflection sequence of the room impulse response as a contributor to timbral discrimination.