Digital communication
In modern techniques, the analog signal is converted into a stream of numbers through a process of sampling, i.e., measuring and recording the signal amplitude at regular intervals. This principle is based on the Nyquist–Shannon theorem, which shows that by sampling a signal at a frequency at least twice its useful bandwidth, it is possible to fully reconstruct it. In the case of voice, typically limited to about 3 kHz, it is therefore necessary to sample at at least 6 kHz.
Sampling produces a data stream that contains, in numerical form, all the information needed to reconstruct the original signal. Once converted into numbers, the signal can be processed using advanced mathematical techniques that allow transmission to be optimized and protected from errors. Unlike in analog systems, the channel no longer directly alters the content of the information, but may instead introduce errors in the numerical data, which can be detected and corrected.
This process is called Pulse Code Modulation (PCM). However, it has an evident limitation: it generates a relatively high data stream. A voice signal with a bandwidth of about 3 kHz, sampled and quantized at 16 bits, produces a data rate on the order of several tens of kilobits per second, continuous even in the absence of speech. Too much data to be transmitted over radio, so it is necessary to reduce the amount of transmitted data without compromising intelligibility.
This is where digital processing comes into play: compression algorithms and error protection techniques are applied to the PCM stream, allowing redundancy to be removed and transmission reliability to be increased.
A useful clarification: in common language the term algorithm is often misused, as if it referred to an autonomous or even intelligent entity. In reality, an algorithm is nothing more than a finite and well-defined sequence of instructions to solve a problem or achieve a result. Even a cooking recipe, described step by step, is in fact an algorithm. Responsibility therefore does not lie with the algorithm itself, but with those who design and apply it.
The algorithms that reduce data volume and make transmission more efficient are called codecs (coder-decoder). A codec analyzes the digital stream and retains only the essential information needed to reconstruct the signal at the receiver, eliminating what is redundant or less significant. In their design, not only the type of signal is considered, but also the characteristics of the transmission medium.
In HF bands (3–30 MHz), for example, typical ionospheric propagation phenomena must be considered: QSB (slow fading), which causes gradual variations in signal strength; selective fading, which affects spectral components unevenly causing distortion; deep fading, which can lead to sudden and severe attenuation; polarization changes due to ionospheric reflections; and ionospheric Doppler, which introduces small but significant frequency shifts, particularly critical in narrowband digital modes. In addition, there is noise, both natural and man-made.
In VHF, UHF, and SHF bands, where ionospheric propagation is negligible, effects related to the local environment and tropospheric propagation prevail. Multipath fading, caused by multiple propagation paths due to reflections from buildings, terrain, or obstacles, generates constructive and destructive interference leading to rapid signal variations and possible reception errors. To this are added Doppler effects due to relative motion between transmitter and receiver (or the presence of satellites), shadowing caused by physical obstacles, and reflections in indoor environments, which can introduce intersymbol interference in digital systems.
Codecs are therefore designed taking these real propagation conditions into account, with the goal of ensuring the best possible performance in terms of intelligibility, robustness, and efficiency. There is no “universal” codec: many exist, often incompatible with each other, each optimized for specific operational scenarios. The overall effectiveness of a digital communication system largely depends on the choice and implementation of the codec.
From channel to information
The real revolution introduced by digital technology lies in the fact that the transmission medium tends to become, within certain limits, transparent to the information being carried. In analog systems, the radio channel directly affects the signal: noise, interference, distortion, and fading phenomena are added to the useful signal, degrading it progressively and irreversibly.
In the digital domain, instead, the channel carries a sequence of bits. As long as these are correctly received-or errors can be detected and corrected-the information can be reconstructed without perceptible degradation. It is no longer the “shape” of the signal that is critical, but the correct interpretation of binary symbols. This represents a profound conceptual shift: communication no longer depends directly on the analog quality of the channel, but on the system’s ability to manage and correct errors.
This abstraction introduces fundamental advantages. On one hand, it enables much more efficient use of the radio spectrum, thanks to compression and coding techniques that allow more information to be transmitted in less bandwidth. On the other hand, it naturally integrates with computer systems, since data transmitted over radio is of the same nature as that processed and routed in digital networks.
It is no coincidence that amateur radio has also played a pioneering role in this field. From early Morse code transmissions-already a basic form of digital communication-through decades of experimentation, increasingly sophisticated systems based on advanced codecs and error correction techniques have been developed. The history of radio can also be seen as a path of progressive convergence toward digital technology: an evolution that has transformed an inherently variable and noisy medium into an infrastructure capable of reliably carrying information and fully integrating into the global telecommunications ecosystem.