0

I am editing this audio clip and I hear a pop at this location.

enter image description here

It might have been cut here, but I am not sure. But regardless of that, I am curious to know whether it looks to you that a pop should be happening here, and the reason why it is happening. The waveform doesn't look too different from other waveforms later in the same clip.

Otherwise, I'd love to know if it's because of my monitors. I just bought some fairly expensive ones and I am a bit annoyed that I am getting so many crackling sounds. Are they normal?

Thank you!

edit: looks like no cut was performed here. it's straight from the mic

4
  • 1
    Please ask the question you need the answer to, rather than add edits & comments. Is it because of your speakers? Clearly not, you can see it in the waveform.
    – Tetsujin
    Nov 25, 2023 at 10:26
  • I can see it in the waveform? No I cannot, as I said, the waveform looks like other sections which don't pop, so why is this waveform popping?
    – Wes
    Nov 25, 2023 at 10:37
  • 2
    If the pop happens at the same place in the file every time you play it back, then the pop is in the file. If it happens at random times that change with each repeat of the same file, then it’s not the file. A common source of pops these days is mismatch of sample rate between two different digital audio devices. Nov 25, 2023 at 15:14
  • With fairly expensive monitors, and if the quality matches the price, you are going to hear a lot of stuff that you previously didn't. Check some other recordings you know, and see if you can hear new things. No recording is perfect. Nov 28, 2023 at 22:25

1 Answer 1

3

Should this "pop"?

YES... That IS a "pop"

If I saw that, yes, I'd expect to hear a "pop" of some kind on that channel, because there's a "sharp" amplitude change, disrupting the continuity of the waveform. Isolated sharp angles like this in a soundwave generally generate a very short burst of broadband noise, whose harmonic content is naturally dependent on the shape and length, as can be seen in this spectrogram of a 440 Hz sine wave with a noisy synthetic "pop" ("synth pop"! ;)).

enter image description here

The nature of soundwaves means any change in amplitude in the time domain will generate additional frequency content in the frequency domain (and vice versa) as we are effectively applying amplitude modulation, and the more abrupt the change, the "brighter" the resulting spectral content. When we build wave shapes by adding sine waves (additive synthesis), we know that in order to create a sharp angle, we must use high frequency harmonics to get the corner as sharp as we can. So when we deconstruct that wave shape via FFT, we know that it will reflect this rich high frequency content. A very useful wave shape in subtractive synthesis is the sawtooth wave, and the reason it's so useful is that it's the wave shape with the richest harmonic content, because it contains these, albeit periodic, abrupt changes in amplitude which could individually be considered "pops" or "clicks".

In reality, as we add constituent waves to build our "pointy" wave shape, the corner becomes increasingly more defined, but as the sample rate limits the amount we can add, they never really become ideal corners. There will always be what we call "ringing" in the analogue signal, as can be seen in this generated sawtooth wave with approximated analogue wave display (using the RX default sample interpolation order of 33). example of "ringing" in a sawtooth waveform

Which brings me to the next problem with "pops" and "clicks". Due to the abrupt change in amplitude, as with the sawtooth wave, a relatively high "true peak" can occur, particularly at higher frequencies. The true peak is the measurement (dBTP) of where the resulting analogue signal is likely to peak.

For example, let's say you purposely create some distortion and think, yeah, it looks fine, the highest peak is around -1 dBFS: Short high frequency wave with about 22 visible samples

But as you can see in the waveform stats, the true peak has a value of over 3 dBTP! That's around a 4 dB increase. If we use the 33rd order interpolation, we get this better approximation of the analogue signal. enter image description here

This will cause more distortion in the resulting analogue signal and could even damage equipment (Although that's not very likely to happen for short bursts like this).

All this results in a relatively intense "pop" or "click" as the intense discontinuous impulse displaces your speaker cone with a fidelity and frequency response dependent on the specs of the monitor/speaker. It disrupts the continuity as if you had given the cone a hard, sharp flick.

So that is why this waveform would make an audible "pop" that may sound different through different output devices. You can see the pop in your spectrogram, though. That doesn't lie. Although it does look like it's not very "sharp"(remember? High frequency content = sharp angle). If there are other similar-looking events, I'd fix them too, with simple interpolation or patching, because they may be a bit smoother than this one, so have a high frequency roll-off, making it a "dull" "pop" that could be masked by other sounds. As for the cause, it could be a sample clock sync "hiccup", possibly a CPU or memory performance issue, as this type of anomalous distortion of the signal is usually the result of a large, approaching instantaneous, step in amplitude.

I hope this is the kind of answer you were looking for.

2
  • Great answer! Now that I think of it and although I haven't tried it some slew rate must be able to help, also, it shouldn't affect the sine wave all that much.
    – frcake
    Feb 7 at 12:00
  • Yeah, there are many ways to fix it and that would probably work in some cases, but I think it's much easier and more generally effective to just select all the samples and use an "interpolate" process, like in iZotope RX. It's like using maths to restore the wave continuity using the surrounding wave. The "quality" slider adjusts the interpolation order, which I suppose is similar to limiting the slew-rate.
    – n00dles
    Feb 7 at 21:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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