Abstract: In a data encoding system, digital data is encoded by adjusting the time position of each transition of a rectangular waveform with respect to the immediately preceding transition and in dependence on the value of the data units to be encoded. In one example disclosed, each data unit is a pair of bits of digital data which can therefore have any one of four binary values: 11, 10, 01 or 00. Immediately after a previous transition (T1), the system encodes the next pair of bits and makes the next transition at a time position (T2) dependent on the value of these bits. If the bits are 00, the next transition T2 is spaced from T1 by eight clock periods. If the bits are 01, T2 can occur either seven or nine clock periods after T1. If the bits are 10, T2 can occur either six or ten clock periods after T1. Finally, if the bits are 11, then T2 can occur either five or eleven clock periods after T1.