What a signal is

Here is a day of room temperature, drawn as a curve. Drag the marker to a few points and read the time and temperature it shows.

Something that changes

Think about the temperature in your room over the course of a day. It’s cool in the morning, warms up in the afternoon, cools off again at night. At any moment you pick, there’s exactly one temperature. Write down that one number, over and over, every minute of the day, and you’ve built what I’ll call a signal: something that changes, with one value at each moment (or position) where you look at it.

That is all a signal is. It is not a formula or a piece of equipment. It is a value that keeps changing, together with what that value was at each moment.

Go back to the curve at the top of the page. Wherever you put the marker, there is exactly one number waiting for you. That is the whole idea of a signal: one value at each place you look.

Not just time

I said “at each moment,” but a signal does not have to be about time at all. Take a photograph and look at just one row of pixels, left to right. At every position along that row, there’s a brightness. Walk along the row and you get a changing value again, just like the temperature, except now the thing that’s changing isn’t the time, it is where you are across the picture.

So the same idea covers a voice recording (a number at every instant of time) and a single row of a photo (a number at every position across the row). Below, you can switch between the two.

Switch between Sound and Picture row and watch the label under the curve change from “time (s)” to “position across the picture (pixels)”. The two curves look different, but you read them the same way: one number for every point along the bottom axis.

One stream or several

All the signals so far have been a single number changing. Sometimes you get more than one number at a time, travelling together. Think of stereo music: one number for what the left speaker should do, another for what the right speaker should do, both changing moment by moment.

So a stereo signal has two values at the same moment, one per channel. Each of those streams of numbers is called a channel. A single stream, like the temperature trace, has one channel. Stereo audio has two.

Switch from One channel to Two channels and look at the Left and Right traces. Notice that they are not a new, more complicated kind of thing. They are two ordinary signals, drawn one above the other, both changing at the same moments in time.

The part you want

Real recordings are rarely this tidy. Say you record a musical tone, but the microphone’s cable adds a faint hiss. What you actually capture is the tone you wanted, mixed together with hiss you didn’t. The part you care about is still the signal; the unwanted part mixed in with it is called noise.

Drag the How much noise slider from 0% up toward 100% and watch the trace. Near 0%, you can clearly see the smooth, regular shape of the tone. As you push the slider higher, the trace gets rougher. At 100% the tone is still visible, but the trace is much rougher.

The next page, Kinds of signals (1.2), looks at other ways to describe signals like this one, including whether they are random or predictable.

From sound to numbers and back

Think about this question: how does your voice, a sound wave in the air, end up as a signal a computer can work with?

Picture a phone call. Your voice reaches a microphone as a wave of changing air pressure. The microphone turns that into a changing electrical voltage, and the phone turns that voltage into a stream of numbers.

Then a computer in the phone works on those numbers. It might remove an echo, or squeeze the list down to fewer numbers to send. At the other phone, the numbers are turned back into sound so the other person can hear you.

Press Next step a few times to move through microphone, turn into numbers, work on the numbers, turn back into sound, and speaker, then press Back to see it in reverse. Notice that most later pages in this course take a closer look at one of these steps: how you turn a sound into numbers, what kinds of work you can do on those numbers, and how you turn them back into sound.

Where you’ll meet signals

Once you start looking for signals, you see them everywhere: the sound of your own voice, a heartbeat trace on a monitor, a room’s temperature over a day, a photograph read out row by row, a radio antenna’s changing voltage. Each one is the same idea: one value at each moment or position you look at.

The next two pages build on this one. Kinds of signals (1.2) gives you a handful of questions to ask about any signal, such as whether you measure it all the time or only at separate moments. How big is a signal (1.3) shows how to put a single number on how big a signal is.

Reference card