An audio-video synchronization method is executable in a video conference device. The method includes determining a first presence time of a predetermined visual effect in a captured video sample stream and a second presence time of a predetermined sound effect in a captured audio sample stream, calculating a time difference between the first and second presence time, and adjusting timestamps of each real-time transport protocol packet in an audio stream sent out by the video conference apparatus based on the time difference. The method further includes receiving an adjustment value from an user input, and adjusting timestamps of each real-time transport protocol packet in an audio stream received by the video conference apparatus based on the adjustment value.