The songs of many songbird species contain trills—rapid repetitions of a single syllable type. Both within and between species, the maximum speed of syllable production (trill rate) is negatively correlated with the trill’s frequency bandwidth, presumably due to biomechanical constraints on the vocal tract. Trills that maximize trill rate relative to bandwidth are said to be high performance, and, in a mate-choice context, female Lincoln’s sparrows (Melospiza lincolnii) prefer male songs that are high performance over those that are low performance. Using Lincoln’s sparrows, we asked whether the trill performance to which males are exposed can serve as a competitive signal and therefore affect their own song behavior. We manipulated songs by cutting silence from between trill syllables of each recording and pasting it in the corresponding space of the recording’s digital copy, thereby making two song-playback treatments—one of low-performance trills and the other of high-performance trills, both within the species’ natural variation of trill performance. In a within-subjects design that was balanced by the order of treatment exposure, we exposed free-ranging males to a 6-minute playback of one treatment one morning followed the next morning by the other treatment. We found no effect of treatment level on the trill performance of subjects during the 10 min immediately following playback. However, relative to the low-performance treatment, the high-performance treatment elevated the trill length and song count of subjects and reduced their songs’ trill-count, syllable-count, and length. These results show that subtle variation in the trill performance to which a male is briefly exposed can significantly alter his song, consistent with the hypothesis that trill performance can serve as a competitive signal between males.