A client for silence-based adaptive real-time voice and video (SAVV) transmission methods and systems, detects the activity of a voice stream of conversational speech and aggressively transmits the corresponding video frames if silence in the sending or receiving voice stream has been detected, and adaptively generates and transmits key frames of the video stream according to characteristics of the conversational speech. In one aspect, a coordination management module generates video frames, segmentation and transmission strategies according to feedback from a voice encoder of the SAVV client and the user's instructions. In another aspect, the coordination management module generates video frames, segmentation and transmission strategies according to feedback from a voice decoder of the SAVV client and the user's instructions. In one example, the coordination management module adaptively generates a key video frame when silence is detected in the receiving voice stream.