Suppose I were to light up a small LED using a potentiometer. It would range from completely off to a maximum voltage or brightness or "loudness".

Guitars works similar. An analog signal carries a certain voltage, but to my understanding voltage on its own only represents the guitars loudness.

What then, is used to represent the actual frequencies of the sound being produced, seeing how voltage appears to only relate to "loudness"?

  • 6
    "What then, is used to represent the actual frequencies": well, the frequency of the electric signal ;).
    – Tom
    Oct 16 '20 at 9:26
  • Be careful about confusing voltage, which is amplitude, with power, which is related to voltage-squared. Oct 16 '20 at 11:54
  • Related, you might enjoy reading Wait But Why's article about what sound (pressure waves) actually is: waitbutwhy.com/2016/03/sound.html
    – JYelton
    Oct 19 '20 at 22:38

While the average RMS voltage represents the average volume, what we hear is our brain's interpretation of the spectral content of the signal. What is that? Let's take a look.

I fired up Audacity, and recorded myself playing an F major triad on a recorder. What do we see when we do that?

F Major Triad

Right away, we can see 4 notes F4, A4, C5, F5, of varying volume. The last note has a larger height, as it required a bit more air and so it was louder. So the thickness of those blobs represents the volume, and if I was pumping this signal out to a speaker, would be proportional to the average voltage. But let's zoom in a bit more on that last note and see what's happening:

F5 note

We can see that F5 isn't just a solid blob, but the signal actually cycles back and forth the whole time. Let's zoom in just a little bit more:

Zoomed in more

This graph essentially shows the movement of air molecules over time. Back and forth, vibrating. It looks a lot like a sine wave. If we measure how many times that peak to peak repetition occurs per second, we'll find out the fundamental frequency. That fundamental frequency is the pitch we hear. In the case of this recording, it's about 697 Hz. F5 for an A440 tuning is 698.46 Hz, so I'm playing just a hair flat.

But, there is more to what we hear than just 697 Hz, after all, we can tell the difference between different instruments. Earlier I said it looks like a sin wave, but it in fact isn't. Or at least, it isn't just a single sine wave. To illustrate, let's see what happens if I pluck a guitar string instead of playing the recorder:

Guitar string

What we can see here, is a dramatically different shape to the repeating signal. We can hear that difference in the waveform's shape, because it has a different spectral content.

Mathematically, every repeating signal can be represented by the sum of an infinite number of perfect sine waves, with varying amplitude at each frequency. This is the spectral content that I mentioned at the beginning of my answer. When we play a note on an instrument, the fundamental frequency will have the largest amplitude, and we'll see large amplitudes at several harmonics, or multiples of the fundamental, and we'll usually see a bit of non-harmonic content as well. The distribution and proportion of that spectral content is what we hear as timbre*. Your brain does this when it interprets the signals from your cochlea. When a sound has a lot of high-frequency components, we hear that as tinny or metallic sounding. Drum sounds often have a much flatter harmonic response with significant non-harmonic components. Electrified instruments can have bizarre spectral profiles which makes distorted guitars and other unique sounds possible that aren't possible with purely acoustic, physical devices. Let's look at some of these in particular.

First the recorder:

Recorder spectral analysis

And the guitar:

Guitar spectrum

Looking at the guitar spectral analysis, we can see significantly more high-frequency components than the recorder, reflective of the metallic sound of its steel strings. The recorder has a much different overtone profile of odd-numbered harmonics, which leads to its "pure" sound. Every instrument will have a different profile on these frequency analysis graphs. Some instruments, the profile will also change from low notes to high notes, in ways beyond just sliding left or right on the horizontal scale, while others will stay relatively consistent.

Lastly, just for fun, let's look at what happens when I play a C Maj chord on the guitar:

enter image description here

What is that? That looks pretty bad. How are we supposed to make sense out of that, where is the the fundamental? Well, the isn't really one, as I'm playing a chord. We have 3 notes speaking, and so we have three of those waveforms piled on top of each other. When some notes are tying to move up, others are moving down, and back and forth, and different frequencies, making that chaotic shape. However, our brains are able to make sense of that and can parse out the frequency components. Let's look at the spectral content:

C Major spectrum

We can see three fundamentals at 264, 330, and 393 for just about C4, E4, and G4 respectively. The overall profile is pretty similar to the single note as well, but we have groups of three for each harmonic, because of each string vibrating.

*So, timber represents the qualitative aspect of the sounds, some of which is shown here by its volume, its pitch, and the spectral content. But there is more to what makes a sound a sound. I was looking at these signals from a steady state perspective, during a sustained note. But the full experience includes the beginning, middle, and end. The sound you hear varies with time after the note is played, (e.g. the sound of the guitar pluck vs the sustain), how quickly the sound dies off, sometimes it changes pitch while sounding out, such as vibrato. There are a lot of aspects that can be modeled, but I think my answer here covers the main aspects.

  • 2
    Absolutely correct! But perhaps a little overwhelming for the level of the question? Oct 17 '20 at 0:22
  • Perhaps, but as someone who is a scientist (even though a computer scientist) and degreed engineer, I strongly appreciated the detail!
    – Astara
    Oct 18 '20 at 17:54
  • As a physics and mathematics nerd I totally enjoyed reading this. My only gripe is the lack of rectangular and sawtooth waves. Oct 19 '20 at 16:27

supposed I were to light up a small LED using a potentiometer, it would range from completely off to a maximum voltage or brightness or "loudness".

Loudness of a sound doesn't relate to how big the voltage is, but across what range the voltage is changing.

Let me put that another way: A very big constant voltage won't relate to 'very loud' - it will represent silence. Sound is all about vibrations and waveforms - pressures (in air) or voltages (in a circuit) that change over time.

Any real physical object that makes a sound does so by vibrating. If we want to represent that vibration in an electric circuit, then the voltage will represent where the object is at a given point in time. A moment later, the voltage will have changed according to where the object has moved to.

So the instantaneous voltage in a circuit does not relate to loudness. However, the range of voltage over a period of time usually does correspond to loudness.

  • 3
    Well, this is correct, but let me add a bit about constant voltage. The power delivered goes as voltage-squared. It's only when we dump current into an electromagnet or other inductive (or capacitative) device that a changing voltage is required to implement motion, said motion leading to a variation in air pressure. I agree that sound is dependent on delta-pressure, so your answer is correct in that sense. A DC current would be more like going into a hyperbaric chamber. Oct 16 '20 at 11:58

Yes, at any given instant an audio signal will have a certain voltage. That voltage would push the cone of a loudspeaker out just SO much. And if that voltage persisted, the cone would stay in that same position and there would be no sound.

But an audio signal exists in time. At THIS instant it has THIS voltage. At the next instant, a different voltage. The speaker cone moves from one position to another, following the ever-changing voltage. Keep changing voltage at short time intervals, it vibrates. And that's sound!


A changing momentaneous voltage can perfectly well represent an (monaural) acoustic signal, just like the changing momentaneous position of a speaker cone (which tends to follow the voltage at the speaker in a certain manner) can represent/reproduce an acoustic signal.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.