A form of communications over a transmission path or channel in which the information is transmitted in digital form. If an analogue signal is to be transmitted in digital form it must first be coded, i.e. transformed into digital form according to some prescribed rules (see digital codes). In such cases regeneration of the original signal requires the reverse coding process – decoding. Digital communication has the advantage over analogue transmission of easier regeneration of the information signal. Significantly more noise and distortion can be introduced by the transmission channel before accurate recovery of the original information becomes impossible. The amount of information that can be transmitted over a path or channel is called the channel capacity.

Information can be transmitted over a channel in a number of different ways. The information signal can be modulated onto a carrier wave by a process of digital modulation. Digital modulation can be achieved using amplitude modulation, in particular amplitude shift keying, pulse code modulation, frequency modulation, in particular frequency shift keying, or phase modulation, in particular phase shift keying. A specific form of digital communication, known as spread spectrum, results when the carrier is itself a digital signal, in particular a repetitive digital code or sequence (see also digital codes). Further variants of spread spectrum include the following: frequency-hopping spread spectrum, in which the frequency of the carrier is periodically or randomly changed; direct-sequence spread spectrum, in which a carrier is first modulated with the information signal, then the information-modulated signal is again modulated with a wideband spreading spectrum signal; transmitted-reference spread spectrum, in which two versions of the spreading carrier are transmitted, one modulated with the information signal and the other unmodulated.

One advantage of digital communications is that, especially in the case of wide-bandwidth channels such as optical fibres, a large number of digital signals can be sent down the same channel; this is usually called multiple access. Multiple access can be achieved in a number of different ways. In time-division multiple access (TDMA), the information sources that need to be transmitted are divided into segments – packets or blocks – and sent down the channel one segment at a time from each source sequentially, or selected by some alternative scheme. In frequency-division multiple access (FDMA), the different information sources are each modulated onto a different frequency carrier. In code-division multiple access (CDMA), the digital carrier code sequence is different for each information source. In carrier-sense multiple access (CSMA), the stations wishing to transmit first listen to determine if anyone else is transmitting, and only if the shared medium is currently not being used do they transmit. A variant on this is called carrier sense multiple access with collision detection (CSMA/CD), in which if a collision is detected, both transmitting stations stop their transmissions and wait for a random period before trying again.

In telecommunications, where a digital message is being transmitted down a switched network it is normally divided into blocks or packets. Each block or packet has a header at the start and a tail at the end: the header contains information about the sender, the recipient, the length of the packet and (optionally) the protocol being used; the tail is generated at the sending end of the communication link by determining the parity characteristics of the message or using a cyclic redundancy check (CRC). At the receiving end the data portion of the message is processed in the same way as at the sending end to allow a comparison of the error-checking parity tail. If the two are the same the message is error-free; if not the block or packet has an error.

There are a number of different ways of dealing with the transmission of blocks or packets of data. If the received block or packet is considered to be error-free, the receiver can transmit an ACK (acknowledgment) signal to the sending end. The transmitting end then sends the next block or packet. This form of transmission is called a stop-and-wait system. If the block has an error, the receiver sends a NAK (negative acknowledgment) signal to the transmitter, which then resends the last block. This process is continuously repeated – ARQ (automatic repeat request) – until an ACK is received.

See also encryption.