An information signal processing method is provided which is to be used for special playback operations to make summary (digest) playback of image/voice data signals of a broadcast program. The method includes the steps of detecting a voice level or predetermined voice characteristic of each predetermined section from a predetermined image/voice information signal or an image/voice information signal resulted from predetermined band compression of the predetermined image/voice information signal and processing the voice signal as a predetermined segment section correspondingly to the result of detection and a predetermined set value; extracting predetermined characteristic data of each section of an image signal from the image/voice information signal and generating predetermined characteristic data indicative of an image feature of the predetermined section from the characteristic data; segmenting the image/voice information signal correspondingly to the image characteristic data, a signal resulted from segmentation of the voice and a predetermined time length or section length setting data; and recording a predetermined voice characteristic signal extracted from the voice signal on the basis of a segment of the image/voice information signal and the image feature data into a predetermined recording medium or predetermined data memory.