A piece of piezoelectric ceramics attached to an aluminum alloy plate was subjected to voltage, and surface displacements of the opposite side were detected by a capacitive displacement sensor. The source waveforms were calculated by a deconvolution integral using Green's function. It was demonstrated that the shape of source waveforms due to the step function of voltage is a triangle, and the rise time of the waveform depends only on the resonance frequency in the thickness direction of ceramics. A wide range of the frequency spectrum of the source waveform was obtained. It was also found that source waveforms due to one period of the sine function of voltage contain three peak values. The shape of waveforms was varied by frequency and phase of applied voltage. Consequently, the artificial AE source using piezoelectric ceramics will be useful for source waveform analysis. Also, it is convenient to prepare learning waveforms for the waveform analysis using a neural network.