This study explores the comparative effectiveness of microphones (standard and ultrasonic) and accelerometers in detecting defects in bearings. Defects in rolling bearings generate acoustic waves with a rich content in high frequencies. Both accelerometer and microphone data allow for Envelope Analysis or Amplitude Demodulation (low frequency), while only ultrasonic microphone data enable high-frequency analysis. The study investigates the advantages of using non-contact microphones to capture this high-frequency content easily. The non-contact nature of microphones also eliminates the risk of altering the machine’s operational conditions, ensuring easy and secure data collection. Various bearings with different defects were investigated by conducting controlled experiments on a test bench. The analysis was carried out by exploiting statistical indicators, e.g., Kurtosis, Crest Factor, Impulse Factor, etc., and energy-based analysis techniques to compare low- and high-frequency results. Results of the different analyses were presented and discussed, showing the potential use of mixed accelerometric, acoustic, and ultrasonic measurements for practical predictive maintenance of rotating machines.