An ultrasound imaging system directs a transmit beam of ultrasound from a plurality of elements in a transducer array into a region of interest (ROI) of a patient's body. The receive beam back from the ROI contains a separate waveform for each of the array elements. These waveforms are partitioned into groups, and a control waveform is determined for each group. The control waveform is then jittered, that is, time-shifted, by a trial delay time, and trial delay times for the other waveforms in the group are determined by interpolation. A waveform similarity factor (WSF), which is preferably a function of the r.m.s. value of the sum of the waveforms in the group, is then evaluated. The control waveform is then repeatedly shifted by different trial amounts, with a new WSF being determined for each trial shift. The trial delay for the control waveform is then assumed to be optimum that yielded the greatest group WSF. A global time compensation profile for the entire array is then determined by interpolation, given the locally optimal time delays of the various control waveforms. This global profile is then applied by a beamformer to compensate the receive beamforming and subsequent transmit beamforming to generate the ultrasound image. The array may also be two-dimensional. The user may select, using input devices and visual feedback, a portion of the displayed ultrasound image to identify a region of interest. The optimum time compensation is then calculated based on the waveforms only in this region, but is applied by the beamformer to the entire displayed image.