There are many analytical and finite element models for predicting vibration of a tire tread. However, experimental verification of these models is limited because of the difficulties involved in measuring vibration near the contact patch of the tire. In this research, a set of experiments were conducted using a microaccelerometer mounted against the tire tread, in the center of the tread pattern. The tire was mounted on Purdue University's tire-pavement test apparatus, a machine that allows precise measurements of tire noise and vibration as the tire rolls over samples of actual pavement. Microphone and accelerometer signals were recorded to determine the influence of pavement parameters on tire-pavement noise generation mechanisms. The vibration measurement and signal processing techniques are verified by comparing the results to published studies. The relationship between vibration characteristics and noise was investigated as a tire rolls over contraction joints in Portland cement concrete pavements. We found that although travelling waves are generated at the leading and trailing edge of the contact patch, the speed coefficients of increased noise and traveling wave amplitude do not match, and the wave speed of the impulse is that of an inefficient radiator. Therefore, increased tread vibration is not a major cause of increased noise at contraction joints, and other mechanisms must be involved.