We investigated vocal adjustment andacoustic niche partitioning in a bird assemblage in the Caatinga based on the Acoustic Niche Hypothesis and the Species Recognition Hypothesis. We conducted 72 recordings of 3-min duration along two transects of 1 km length, between July 2022 and February 2023. Each transect was divided into six points, spaced 200 m apart. We analysed the spectral and temporal overlap (Pianka index) of 15 vocally dominant species and tested significance using null models based on 1000 repetitions of the Randomization Algorithm 3. Phylogenetic distances were estimated from 2500 phylogenies to assess their correlation with acoustic overlap, using Generalised Linear Models and a partial Mantel test. Twenty-five species pairs exhibited high spectral overlap (> 0.6), whereas all pairs showed low temporal overlap (< 0.4). The analysis with null models suggested that the observed spectral overlap (0.322) was lower than expected by chance (0.261), indicating a significant difference [Standardised Effect Size (SES) = 7.162, p < 1]. The temporal overlap showed low values, with the observed (0.097) being slightly lower than the expected (0.111), but with no significant difference (SES = -6.752, p > 1). Our results revealed no correlation between spectral and temporal overlap (Estimate = -0.0059, t = -0.286, p = 0.766) or between phylogenetic distance and spectral (Estimate = -0.0001, t = -0.966, p = 0.336) and temporal (Estimate = 0.00002, t = 0.125, p = 0.901) overlap. There was also no significant correlation between spectral and temporal overlap controlled by phylogenetic distance (r = -0.02712, p = 0.553). There was no evidence of a negative association between the temporal and spectral overlap among the Caatinga bird species. The phylogenetic relationships did not explain the observed patterns. Therefore, it is possible that temporal avoidance actively occurs regardless of other factors, such as spectral overlap.