A videophone system providing high resolution video transmission between videophones utilizes compressed video signals and audio signals which may be transmitted through any communications network, with the system providing real time adaptive error recovery and synchronization between the audio and video signals to produce high quality video images and sound. The system is highly resilient, adapting in real time to changing conditions in a network or to network errors due to data corruption or loss that can be produced, for example, by noise or line losses, thereby substantially eliminating fitter, signal packet loss or delay, or other errors which produce signal degradation.