PURPOSE: To obtain a voice together with a character picture without a waiting time, by storing a voice data in a multi-page memory, and also synthesizing the voice data in response to read-out of the voice data.

CONSTITUTION: As for a video signal on which a character broadcast is superposed, a character signal is separated from a TV signal by a sampling circuit 4 and is stored in a buffer memory 5 temporarily. A signal of a program selected by a program selecting device 10, out of the contents of said memory is data- processed by a CPU6, and a character pattern data, a color data and a voice data are transferred to a pattern memory 7, a color memory 8 and a voice memory 20, respectively. Subsequently, the character pattern data of the memory 7 and the color data of the memory 8 are read out successively by a read-out controlling circuit 11, and when they are sent to a CRT13 through an output interface circuit 12, a character picture is obtained on a TV picture. Also, the data of the memory 20 is outputted in a voice 18 through a decoding circuit 22.