In telecommunications, frame synchronization or framing is the process by which, while receiving a stream of fixed-length frames, the receiver identifies the frame boundaries, permitting the data bits within the frame to be extracted for decoding or retransmission.

When packets of varying length are sent, it is necessary to have an instantly recognizable packet-end delimiter (e.g., Ethernet's end of stream symbol). Loss of carrier signal can be interpreted as a packet-end delimiter in some cases. When a continuous stream of fixed-length frames is sent, a synchronized receiver can in principle identify frame boundaries forever. In practice, receivers can usually maintain synchronization despite transmission errors; bit slips are much rarer than bit errors. Thus, it is acceptable to use a much smaller frame boundary marker, at the expense of a lengthier process to establish synchronization in the first place.

Frame synchronization is achieved when the incoming frame alignment signals are identified (that is, distinguished from data bits), permitting the data bits within the frame to be extracted for decoding or retransmission.