Bearings are critical mechanical components that are used in rotary machinery. Timely detection of defects in such components can prevent catastrophic failure. Noise is generated during the rotation of bearings even without the presence of defects due to finite number of rotating elements to carry the load. Such noise is associated with the change in effective stiffness during rotation, however, a sharp spike is observed in the noise level with presence of local defects. This study uses the noise generation aspect of roller bearings to identify local defect in a single row ball bearing with outer race stationary under radial load. Experimental testing is conducted on two identical bearings. The defective bearing is selected from a diesel engine subjected to 20 years of service. Dissecting the defective bearing revealed pitting and spalling of the inner race and balls, the most two common bearing defects. Both time and frequency analysis of sound pressure generated by the bearings were performed. The results show that there is a clear distinction in the time and frequency spectra between healthy and defective bearings. Findings of this study revealed that using a simple cost efficient in-house experimental setup, local defects can be readily detected.