Abstract: An audio analysis system visually displays a variance of sound intensity in each of several discreet frequency bands from the average sound intensity across all frequency bands. The audible sound is detected and converted to a first electrical signal which represents the ambient sound detected. A second electrical signal is developed as a function as the average DC level of the first signal. The DC level of the first signal is then shifted in accordance with the average level of the second signal to develop a level shifted signal. The level shifted signal is then graphically displayed in the discrete frequency bands so that the plus or minus deviation from the average level signal is displayed in each of those bands. The display may be provided with a visual representation, such as a graphic overlay or multicolor display, showing the probability of risk to hearing loss produced in each of those bands.