There is an increasing interest in the use of autonomous recording units as an alternative survey method to point count surveys conducted by human observers; however, questions remain about whether or not these recording units perform similarly to point count surveys and produce valid, comparable results. The use of individual listeners to transcribe the acoustic data collected by autonomous recording units is a common method for the analysis of recorded acoustic data, but potential variability among transcribers raises questions about the standardization of listening protocols to decrease inconsistencies in transcription results. Autonomous recording units have been used to monitor birds in and around Brimfield State Forest in Brimfield, Massachusetts since 2012, after a tornado severely damaged a large area of the forest and surrounding properties. In 2016 and 2017, I conducted 71 10-minute point count surveys while simultaneously recording the survey with an autonomous recording unit in three habitat types in and around Brimfield State Forest in Brimfield, Massachusetts. I transcribed the acoustic data from the recordings and compared it to the results of the point count surveys to determine if autonomous recording units performed as well as point count surveys. To assess variability among listeners, four listeners transcribed the same sample of 30 recordings and a sub-sample of 6 of those recordings that were created during the 2014 field season using two different listening protocols. The first protocol instructed listeners to play each recording straight through without stopping, and the second protocol instructed listeners to stop and replay any part of the recording they needed to and also use outside sources to aid in species identification. I compared the number of species, individuals, distant individuals, and mean counts (uncorrected abundance), corrected abundance and detectability of focal species between both survey methods, among all listeners using both listening protocols, and where possible between habitat types to assess differences in method performance and listener variability. I tested for correlation between autonomous recording units and point counts using the uncorrected and corrected abundance estimates. The number of species and number of individuals detected did not differ between survey methods overall and for each habitat individually; however, in each habitat type, more individuals were classified as distant by autonomous recording units overall for all habitats. The number of species detected did not differ between listeners overall and for in each habitat using either listening protocol. The number of individuals and distant individuals detected differed significantly between listeners and within certain habitats using the first listening protocol. There were no differences in the number of species, individuals, or distant individuals detected overall between listeners using the second listening protocol, but there were significant differences in individuals and distant individuals detected between habitats by listeners. Corrected and uncorrected abundance estimates between autonomous recording units and point count surveys were highly correlated, and there were no differences in detection probabilities for the focal 23 species between survey methods and among habitat types. Only 2 out of 18 focal species indicated a significant difference in detection probability between listeners using both listening protocols. Based on the results of my study, I conclude that autonomous recording units perform at least as well as human observers conducting point count surveys, and that multiple listeners transcribing the same acoustic data do not show high levels of variation in the results of their transcriptions.