A dynamic process of perceiving a frequency-modulated (FM) tone has been analyzed. A warped pitch image perceived for a linear sweep and pitch ringing induced by a piecewise linear sweep suggested the existence of a dynamic FM tracking process. The dynamic characteristics of the tracking process are represented by a second-order system. To track multiple FM tones, a neural matrix model was proposed based on the second-order system model. Several interesting perceptual effects were predicted by this model. To further analyze the dynamic process, perceived pitch images for ascending or descending pure-tone sequences were newly examined. A stimulus is a sequence of 20 pure tones. The total duration of the sequence is 632 ms. Rise and fall times of each pure tone are 5 ms each. When the frequency ascends linearly in the logarithmic frequency scale, the frequencies of two adjacent tones are separated by a semitone. Subjective tests revealed that a sequence of pure tones is perceived as a warped pitch sequence. All results are similar to the results obtained for continuous FM tones. The results suggested that the global pitch image of a sequence of pure tones is produced by the dynamic processing of the tone sequence.