Difference between analog signal and digital signal

Question

There is no perfect generator of a true digital signal in the real world natural phenomenon. Most of the physical signals produced by humans and animals like roaring, speaking, playing, clapping, electrostatic charging, drumming are all analog in nature. Even physical heavenly bodies like sun, earth, planets, moon, rocks, mountains generate signals that are analog in nature like seashore sound waves, light pattern, radioactive decay, gravitational pull, centrifugal force, natural growth rate are all analog in nature.

Correct me if I'm wrong, what i see is that all naturally generated signals are analog in nature. What we see as digital signals are those analog signals sampled in space or time. Is there exists a true generator source of digital signal in nature? Is it correct to say that analog signals are natural and digital signals are man-made?

Answer 1

The practice for calling signals digital or analog in communications systems is that analog signals carry information in continuous or semi-continuous changes in the signal, whereas digital ones carry information in discrete changes. Discrete and digital are used interchangeably. See for instance Stallings basic books, or see modern and excellent books in communications like Proakis and Salehi Communications Systems Engineering. Wikipedia may also be ok, but they're never really correct (I teach a course on it and have to always warn the students). Since the information carried is the relevant factor, the provisioning of information to some electromagnetic (or other) wave (also called the carrier wave, or carrier) is called modulation, or sometimes more generally encoding.

Wiki is at https://en.wikipedia.org/wiki/Digital_signal. No guarantees.

After the modulation everything, carrier and modulation, ie, the full signal, is usually upconverted to the frequency at which it is meant to be transmitted, sent tot he antenna, and radiated away. Yes, of course, the actuial physical wave sent out, with the carrier imparted with the modulation (ie, the changes), is analog, in the sense that even the discrete changes take a small amount of time, and the power being sent does have some, typically very small, power in the time during which is 'discretely' changes.

Thus an FM modulated signal is analog, also AM and PM, where the changes are continuous. The digital (discrete) equivalents are FSK, ASK and PSK, where the freq, amplitude or phase changes quickly, called discontinuously (more on this in next paragraph, it always takes a bit of time as stated above). In digital signaling there is a lot lot lot more control and precision one can exercise over the signal features, and thus more information. Digital circuits also make it much cheaper.

As for those discontinuities, as you probably know from Fourier transforms and from quantum mechanics, any fast change in the time domain of the signal, called the waveform to emphasize it's changing shape, will mean the time uncertainty of the change is small, and thus the energy will be spread over a larger range, creating what is interference in nearby frequency bands. Thus part of the trick (now a well understood set of tradeoffs) is to have some shaping of the transition for the optimal tradeoff desired. In modern LTE 4G wireless comms the carriers are separated via an inverse Fourier transform so that the interference with adjacent frequencies is theoretically zero, at the specific transition time, but again, still some design tradeoffs.

So, yes if you are a physicist or an RF engineer or technician you might say all signals are analog. But the practice in communications engineering (systems and design engineers, practitioners and technicians) is to call those with very rapid transitions, digital.

Answer 2

The difference between an analog and a digital signal is in how the receiver treats it. (At least, that's one definition.) In an analog signal, any value of that signal goes, but therefore the receiver can not know what the original sent value was and what the contribution of noise was to the received value. In a digital signal, the receiver assumes that the sent signal was one of a small set of discrete values, and if the received signal value deviates from one of this set of values, the closest value was what the sender sent and the difference was noise. I find this answer to have a good description of the basic difference.

For a receiver it only makes sense to treat a signal as digital if the receiver can make the assumption that the signal sent was from a finite set of discrete values. One reason such an assumption can be made is if there is an agreement with the sender to do that. In most cases such an assumption can not be made for natural signals, so most of the time they can only be treated as analog.

Humans and animals can have such agreements, either explicit or by instinct. Human language can be said to be partially digital. We usually interpret speech from someone else as being in the known vocabulary of the language he's speaking, even if the actual sounds don't faithfully match a clear pronunciation of a word. This mechanism usually works unconsciously without us being aware of it.

An exception would be someone introducing him or herself with a name you have never heard of. It can take a few tries before you correctly get the name. To do that you need to drop down to the lower level of the discrete set of sounds that are used in your language. If the new name is from another language that uses different sounds, you either need to learn that new sound, or (what usually happens) people will pronounce the name wrongly, using the closest sounds in their own language. Either way this new name forms a new item in the agreed discrete set of word values you use in your conversation.

Of course other aspects of language, such as tone, mode, voice, etc. are not digital.

The same can be said for some animal communication. Some animals can make warning signals when e.g. a predator is nearby. Such signals will be treated 'digitally' by other animals in the group: no signal or danger!

If you dig down deep enough you can find some digital aspects in many kinds of signals, at least if those signals are sent intentionally. The mere fact that there is a signal or not is usually digital. We can safely assume that an analog TV transmitter station is either trying to send a signal, or not. It is practically always safe to exclude the possibility that the tv station is sending its signal at such a low power that it is barely detectable. If we receive such a low power signal we can assume that that is because of e.g. the long distance to the transmitter or the weather, but not due to the transmitter sending at very low power.

We can also sometimes make some discreteness assumptions for natural phenomena based on our knowledge about them. For example, while many aspects of celestial objects are analog, the fact that an object exists is not (in most cases, for compact objects). So even detecting a very weak signal with a telescope means that an object has been detected.

Answer 3

There is a sense in which everything is quantized, therefore an analog signal is really a digital signal in disguise.

If you take a very sensitive detector of say a radio signal and then turn the signal down low enough, you will begin to see the signal strength varying in small steps that match the energy of one photon at that frequency. Turn the signal down even lower and you will see it arriving in discrete jumps, one photon at a time. If the signal is unintended this is known as shot noise and can plague the design of the most sensitive detectors.

Just as say a 16-bit digital audio channel sounds analog to our ears, so the world is built of digits and only appears analog.

But this is not the whole story, for the exact energy of a photon, or say the kinetic energy of an electron, may take any value along a wide analog scale.

A friend of mine once coined the term "digalog" to describe such mixed analog and digital properties (which he incorporated into his measuring instruments), but I have not seen it widely used elsewhere.

Answer 4

You are right. Everything in nature is analog (including man-made "digital" signals). The digitization of signals is a man-made abstraction. Nature just evolves as it does and humans then categorize observed phenomena. The assigning of one signal as a "1" and the other signal as a "0" is a categorization.

It is possible in nature to have a system that has bimodal output (that is outputs are clustered around two qualitatively different behaviors) but there is always some noise and everything is "analog" in the end.