A method and an arrangement maintain end-to-end synchronization on a telecommunications connection transmitting data in frames substantially in real time and using synchronized end-to-end encryption, wherein at least a part of the telecommunications connection is a packet-switched connection, in which case the reproduction delay of the data to be transmitted can be increased by adding one or more extra frames to the frame string being transmitted, wherein the arrangement defines, based on the number of received frames, an initialization vector value corresponding to a frame received at the receiving end of the telecommunications connection and used in decrypting the frame, adjusts the reproduction delay to mark the frame to be added to increase the reproduction delay as an extra frame, and defines the initialization vector value to count only the frames not marked as extra frames in the number of received frames.