Analogue and Digital Signals
Every signal you'll meet in this topic — a sound wave from a microphone, a temperature reading from a sensor, a stream of data down a fibre-optic cable — falls into one of two fundamentally different categories. Understanding the difference between them is the foundation on which the rest of this module is built.
Analogue signals: continuous quantities
An analogue signal is a quantity that varies continuously over a range of values, able to take any value at all within that range. If you plotted an analogue voltage against time, you'd get a smooth, unbroken curve — it might rise, fall, oscillate or wander, but at every instant it has some definite value, and as time changes by an infinitesimally small amount, the signal's value also only changes by an infinitesimally small amount. There are no jumps, no gaps, no forbidden values in between.
This matters because almost every physical quantity in the natural world is inherently analogue. Sound pressure varies smoothly as a wave passes through the air; light intensity varies smoothly as a source brightens or dims; temperature varies smoothly as a room warms up. None of these quantities naturally come in neat, separated steps — they take every value in between as they change. Consequently, a sensor built to measure one of these quantities — a microphone, a light-dependent resistor, a thermistor — typically produces an analogue output: a voltage or current that varies continuously, in proportion to whatever physical quantity is being measured. A thermistor's resistance (and hence, in a suitable circuit, its output voltage) drifts smoothly as temperature drifts smoothly; a microphone's output voltage traces out a smooth waveform that mirrors the smooth variation in air pressure reaching its diaphragm.
Digital signals: discrete states
A digital signal, by contrast, is restricted to only a small number of discrete, distinct states. In electronics this is almost always just two states, conventionally described as "high" and "low", or represented numerically as 1 and 0. A digital voltage doesn't smoothly track every intermediate value the way an analogue one does — it is designed and intended to sit at (or very near) one of just two levels, for example close to 0 V for "low" and close to 5 V for "high", with values in between treated as invalid, transient, or simply not meaningful as data, rather than as genuine intermediate readings.
This restriction to a handful of discrete levels is what gives digital signals their name, and it is also the single property from which almost every practical advantage of digital electronics ultimately flows — a point this lesson returns to at the end, and one that's developed much further in the "Noise and Signal Recovery" lesson later in this topic.
Bits and bytes
Because a digital signal has only two possible states, each one can be used to represent a single binary digit, or bit — a 0 or a 1. A bit is the smallest possible unit of digital information: on its own it can distinguish between only two possibilities.
To represent more than two possibilities, bits are grouped together. A group of 8 bits is conventionally treated as a single unit of data called a byte. Because each of the 8 bits can independently be either 0 or 1, the total number of distinct patterns (and therefore the number of distinct values a byte can represent) is $2^8 = 256$. More generally, a group of $n$ bits can represent $2^n$ distinct values, since each additional bit doubles the number of combinations available.
Where sensors fit in
This gives the practical starting point for the whole of this topic. Sensors collect data from the physical world — temperature, light intensity, sound pressure, and so on — and because these are inherently analogue quantities, the raw output from a sensor is normally an analogue voltage that varies continuously in proportion to whatever is being measured.
Yet almost all modern processing, storage and transmission of data — computers, digital audio, mobile networks, the internet — is built around digital signals, not analogue ones. This creates an obvious problem: how do you take a continuously-varying analogue voltage from a sensor and turn it into a sequence of discrete 0s and 1s that a digital system can work with? Answering that question — the process of analogue-to-digital conversion — is the subject of the next lesson.
Given that the real, physical world is fundamentally analogue, it might seem odd that so much modern technology converts everything to digital as quickly as possible. The key reason is robustness against noise and degradation. A digital signal only ever needs to be read as one of two levels — so if some noise or interference slightly corrupts a "high" signal, a receiving circuit can still correctly read it as "high", as long as the corruption hasn't pushed it past the halfway point toward "low". An analogue signal has no such safety margin: any noise added to it changes its exact value, and there is no way for anything downstream to tell how much of the received value is genuine signal and how much is corruption. This idea is developed fully, with the reasoning behind signal recovery, in the "Noise and Signal Recovery" lesson later in this topic.
Key takeaways
- An analogue signal varies continuously and can take any value within a range; a digital signal is restricted to a small number of discrete states, almost always just two ("high"/1 and "low"/0).
- Real physical quantities — sound pressure, light intensity, temperature — are inherently analogue, and sensors that measure them typically produce a continuously varying analogue output.
- A bit (binary digit) is a single 0 or 1, the smallest unit of digital information.
- A byte is a group of 8 bits, capable of representing $2^8=256$ distinct values.
- More generally, $n$ bits can represent $2^n$ distinct values, since each extra bit doubles the number of possible combinations.
- Sensors collect analogue data from the physical world, but modern processing, storage and transmission systems are overwhelmingly digital, creating the need for analogue-to-digital conversion (covered next).
- Digital signals are generally preferred because they are far more robust against noise: a corrupted digital signal can often still be read correctly, whereas a corrupted analogue signal's true value is simply lost.