Abstract: This paper introduces a method of extracting the useful speech onoff signal based on the features of conversation speech, and the frame structures of implement voice/data integration in a single channel employing the speech interpolation technique.