Multiplexing is the process of combining multiple data streams into a single shared channel for transmission. Demultiplexing is the reverse: separating that combined signal back into its original individual streams at the receiving end. Together, they allow a single communication link to carry information from many sources simultaneously, which is how everything from phone networks to internet connections to satellite telemetry handles enormous volumes of data without needing a dedicated wire or frequency for every single user.
A multiplexer (MUX) sits at the sending side and merges the signals. A demultiplexer (DEMUX) sits at the receiving side and sorts them back out. The two always work as a pair.
How Frequency Division Multiplexing Works
Frequency Division Multiplexing (FDM) is one of the oldest and most intuitive approaches. Each data source gets its own slice of the available frequency range. Think of it like radio stations: dozens of stations broadcast simultaneously, but each one occupies a different frequency band, so your radio can tune into just one without hearing the others.
Each source’s signal is shifted onto a different carrier frequency before being combined. To prevent signals in neighboring bands from bleeding into each other, small unused gaps called guard bands are placed between them. These guard bands act as buffers, absorbing any slight frequency drift so channels stay cleanly separated. The demultiplexer on the receiving end uses filters to isolate each frequency band and recover the original signals. FDM is widely used in cable television, FM radio, and traditional telephone trunk lines.
How Time Division Multiplexing Works
Time Division Multiplexing (TDM) takes a different approach: instead of splitting frequency space, it splits time. The shared channel is divided into repeating time slots, and each source takes turns transmitting during its assigned slot. It comes in two flavors.
Synchronous TDM
In synchronous TDM, time is divided into fixed periods, and each period is split into a set number of slots. Every user gets the same slot position in every period, guaranteed. This predictability is its strength: a waiting message will always be transmitted during the next period, so you get reliable, consistent timing. The downside is waste. If a user has nothing to send during their slot, that slot goes empty. No one else can use it, because it’s reserved for that specific user regardless of whether they need it.
Asynchronous (Statistical) TDM
Asynchronous TDM, sometimes called statistical multiplexing, eliminates that waste by letting any active user grab the next available slot. Slots are only empty when every single user is idle. This makes far more efficient use of the channel’s capacity, since bandwidth isn’t sitting unused while one source has nothing to send. The tradeoff is that you lose timing guarantees. How quickly a user gets a slot depends on how many other users are active at that moment, so latency can vary unpredictably under heavy load.
The efficiency difference matters at scale. Research from the University of Texas has shown that intelligently managing how different traffic types share bandwidth through statistical multiplexing can reduce blocking (the chance a transmission gets denied) by an order of magnitude compared to simpler allocation methods.
How Code Division Multiplexing Works
Code Division Multiplexing (CDM) allows multiple users to transmit on the same frequencies at the same time. Instead of separating signals by frequency or time, it separates them by assigning each user a unique digital code. Each sender encodes their data using their signature code and broadcasts it into the shared medium. The receiver knows the intended sender’s code and uses it to extract just that signal, treating everything else as background noise.
This technique became widely known through CDMA cellular networks. It’s particularly useful in wireless systems because it’s resistant to interference and can handle many simultaneous users across the same spectrum. Where FDM uses different frequencies like different radio stations, CDM uses different codes like different languages being spoken in the same room. You can pick out the one you understand and ignore the rest.
Multiplexing in Modern Wireless Networks
4G LTE and 5G networks rely on a technique called Orthogonal Frequency Division Multiplexing (OFDM). It divides a wide frequency band into many narrow subcarriers spaced precisely so they don’t interfere with each other. The “orthogonal” part means each subcarrier’s peak lines up exactly with the nulls of its neighbors, allowing them to overlap slightly without crosstalk. This packs more data into less spectrum than traditional FDM.
4G LTE uses a fixed subcarrier spacing of 15 kHz. 5G is more flexible, defining five different configurations with varying subcarrier widths. Wider spacing helps at higher frequency bands because it resists the signal distortion that comes with fast-moving devices. It also enables shorter time slots, which means faster packet delivery for applications that need ultra-low latency, like remote surgery or autonomous vehicle communication. This flexibility is a core reason 5G can serve such diverse use cases, from massive sensor networks to high-speed mobile video.
Multiplexing in Computer Networks
Multiplexing and demultiplexing aren’t just physical-layer concepts. They happen at the software level too, particularly in the transport layer of computer networks (where TCP and UDP operate).
When your computer runs a web browser, a video call, and a file download simultaneously, all that data flows through the same network connection. The transport layer handles multiplexing by taking data from multiple application sockets, tagging each chunk with a destination port number, and packaging it into segments that get sent over the network. On the receiving end, demultiplexing works in reverse: the transport layer examines the port number on each incoming segment and delivers the data to the correct application socket. Port 443 goes to your browser, port 53 goes to your DNS resolver, and so on.
This is why you can have dozens of applications using the internet simultaneously over a single network interface. Each one is identified by its port number, and the transport layer sorts everything to the right place.
Real-World Applications
Multiplexing is embedded in technology you interact with daily, and in specialized systems you might not think about:
- Cable and broadcast television: Hundreds of channels share a single coaxial cable or broadcast spectrum using frequency division multiplexing. Your set-top box demultiplexes the signal to display only the channel you selected.
- Telephone networks: Traditional phone systems used TDM to carry thousands of voice calls over a single fiber or copper trunk line, giving each call a repeating time slot.
- Satellite telemetry: Orbiting satellites transmit dozens of sensor readings (temperature, radiation, position) over a single radio link using multiplexed channels. Ground stations demultiplex the signal to separate each measurement.
- Oceanographic monitoring: Unmanned instrumented buoys in the ocean are interrogated by a central station at set intervals, with multiplexed channels carrying data on wave height, water temperature, salinity, and other variables over a single radio link.
- Biomedical research: Miniature transmitters swallowed or surgically implanted in patients send biological measurements from inside the body. Multiplexing allows multiple data streams (heart rate, pH, pressure) to share the tiny transmitter’s limited bandwidth.
Hardware: MUX and DEMUX Circuits
At the hardware level, multiplexers and demultiplexers are fundamental digital logic components. A MUX chip has multiple input lines, a single output line, and a set of selection lines that determine which input gets routed to the output. The relationship follows a simple rule: with n selection lines, you can choose from 2^n inputs. So 2 selection lines handle 4 inputs, 3 selection lines handle 8, and 4 selection lines handle 16.
A DEMUX works in reverse: one input, multiple outputs, and selection lines that determine which output receives the signal. These chips are building blocks in everything from computer processors to digital displays, routing data through circuits the way a railroad switch routes trains onto different tracks.