To surely recognize call contents and to enhance the compression rate of data to be recorded by converting digital voice data of inputted call signals into text data and recording them in an information recording medium as recording data.

An A/D converting part 3 converts the analog call signals of plural channels into digital voice data. Digital voice data are converted into text data at every channel by a voice recognizing part 5. A writing converting part 7 generates head data concerning text data and adopts them as recording data consisting of head data and text data. A recording and reproducing part 8 records recording data in the information recording medium. Therefore, a recording information quantity is reduced and long-term recording is enabled. Text data are voice-converted and recognition is executed by synthetic voice. The call signals can be heard clearly by the same recording information quantity compared with a case when digital compression data with the hightened compression rate are reproduced so that the call contents are surely recognized.