A critical step in mechanism noise reduction for data-processing equipment is identification of sources. Techniques such as disabling mechanisms, microphone probing, and frequency analysis are useful here. We have discussed procedures based on noise-time analysis within machine cycles at past Acoustical Society of America meetings as being particularly useful for this work. For high-speed impacting machinery, normally used analog acoustical instruments are inadequate for noise-time analysis. Procedures have been devised using a high-speed electronic switch attenuator in a “stroboscopic” mode to overcome these limitations. This technique has been useful for source identification and examples are presented. Digital procedures have considerable potential for these applications. Fast Fourier transform digital filtering, and related techniques are useful and we have used these procedures where applicable in noise-time analysis using an IBM 1130 computer system. The practical limit here is in processing time and/or computer size (and cost). Our present direction is combining the best of both digital and analog techniques in a hybrid system. The resulting approaches will be discussed as more nearly optimum than other approaches and most usable for mechanism noise identification.