A voice track is recorded on one track of a dual track recording device, typically a tape recorder. On a second track of the recorder an animation signal is recorded characterized by the signal having a frequency at any given time indicative of the then currently desired animation condition. A character is provided having a tape playback unit therein for playing back such pre-recorded dual track tapes, the character having an amplifier and speaker for reproducing the audio information, and a servo motor having a drive system for driving the character mouth elements and eyes with the desired animation, the drive system having a feedback device thereon for providing a feedback signal to the servo motor control. The electronics in the character which is responsive to the animation control signal provided by the playback unit to provide the servo motor drive signal effectively updates the servo motor drive signal on each cycle of the animation control signal received from the playback unit.