The term “digital” is in some ways confusing. Yes, digital signals depend on numbers to work, but I’m going to call them “descriptive signals”. You’ll see why in a bit.
When you send an analog signal, you build a model of the thing you want to send. Usually the form differs, but you can still see the original if you know how to look at it. A landline phone is the classic example. You speak. A microphone picks up the sound waves and converts them to a fluctuating voltage. But if you plot out that voltage over time on a graph, it looks like the sound wave did. This signal goes to a speaker on the other end, which can convert that fluctuating voltage into sound waves, and since the voltage looks like the sound waves for wbat you said, the speaker outputs something that sounds like what you said. Let’s say that instead of going to a speaker, you fed the signal to a vibrating vinyl cutter, for making a record. It traces a groove into the disc, but if you look at those grooves in a microscope, they *also* look like the sound wave. This is the key to understanding analog signals: at every step, you’re creating a model of the previous step, just in a different form.
Digital signals -or rather, descriptive signals- work in a fundamentally different manner. Rather than making an entire model of the thing you want to transmit, you describe it in some agreed-upon manner. When you make a digital recording, a computer analyzes the sounds you make and builds up a description of the waveform. A player then takes this description and builds a waveform that fits that description. How good the new sound is depends on how well the original computer was able to describe it.
Which is better? In some ways, that depends on what you want to do. Descriptive signals are typically easier and less error-prone to copy. Anything that interferes with an analog signal becomes part of the signal, and can be very difficult to remove. With descriptive signals, only the things you’ve defined to be part of the description actually matter, so interference is usually easier to spot, and then you can ignore it or correct it. Descriptive signals are rarely perfect -descriptions can only go into so much detail- but because of the interference problem, analog signals aren’t usually perfect either. The interference problem becomes a big issue for copying analog signals, because imperfections in the machinery add up over time, while those imperfections generally don’t matter for descriptive signals anyway.
But my favorite example, sadly, may be out of reach now. See if your school’s library still has any microfiche equipment (it probably doesn’t, but check). This was popular for storing periodicals like newspapers and magazines before the advent of mass storage like CD-ROM. Newspapers are not yet so obscure that your students wouldn’t have seen them, and you can imagine how much room it would take to store an entire year’s worth of them. How did libraries handle these huge numbers of periodicals, back before the invention of digital storage? The analog solution was to literally make the periodicals smaller: they took photographs on these plastic sheets and shrunk them down very small. Microfiche readers are essentially small projectelors that enlarge the image so humans can read them, but you also read them with a microscipe if you really had to. This is the analog form of compression: to shrink the signal, make a model that’s actually smaller (in whatever way you need it to be smaller).
Digital -or rather, descriptive compression- works differently. You have to somehow make the description more concise. You can decide that some things don’t matter and just throw those pieces away: this is *lossy encoding*, because anything you throw away is just lost. As long as you’re judicious about what you throw away, this can be very efficient, but you have to be careful not to throw anything that’s actually important, or people will notice. Or, you can try to find a more efficient way to say your description, without actually throwing anything out. For example, if some long string gets repeated a few times, you might replace them with a short code that means the longer string. You also have to store a dictionary for these codes, but if the longer strings are repeated many times, you can still save space overall. This is *lossless compression*.
Latest Answers