Audio Distortion


I have just recently started making some youtube videos and after I uploaded them, I noticed a very loud distortion / crackling when the audio went too high. I’m kind of new to the audio thing, but the sound went into the red.

Is there anyway I can avoid this?

A sample is here:

I know it was because my laugh was too loud. Note that I wasn’t using audacity to record, but the interface through my HD PVR2 used to record gaming footage. Any insight or help would be appreciated. I have done some searching and it looks like this might be “clipping” or “peaking”.

Is there anyway I can avoid this?

Don’t laugh? Seriously, my laugh will go through soundproof walls.
Both digital and analog recording systems “run out” if the sound is too loud. Many analog systems overload gently, but digital systems crack and crash like what you have.

In Audacity you have to pay attention to the recording meters which is why they’re much bigger in the latest versions. Some recording systems have auto limiters to prevent this.

Your mileage may vary, consult your local listings.


Would setting normalize to -6 or -1 maybe help? Sitting farther away from the mic?

Would setting normalize to -6 or -1 maybe help?

No. Normalizing or adjusting the volume after recording doesn’t change the shape of the waveform. If the waveform is [u]clipped[/u], or if your ADC (analog-to-digital converter) is driven to clipping, there is nothing you can do.

Sitting farther away from the mic?

Yes. Backing-off the mic, or if you have an analog gain control you can turn it down.

It’s a balancing act to get a good strong signal without overload/clipping and a low noise level (for a good signal-to-noise ratio). It’s difficult to get a professional sounding recording “at home” because most of us don’t have soundproof studios or pro-quality low noise electronics. (Although for a couple-hundred dollars you can get a near-professional quality audio interface.) A directional mic helps too, since room noise comes from all directions and the signal comes from only one direction.

Backing-off the mic or will reduce the signal without reducing room noise or electrical noise. (Worse signal-to-noise ratio.) Turning-down the gain analog will reduce the signal, room noise, and some electrical noise, but there will still be some remaining electrical noise (from the analog circuitry following the gain control or from the ADC itself).

After recording, I check the recording by running the Amplify effect. The Amplify effect scans the file and defaults to whatever gain or attenuation is required for 0dB normalized/maximized peaks. For example - If immediately after recording*, Amplify defaulsts to +2dB (gain), your current peaks are -2dB (digital signal level) and you have 2dB of “headroom”. If Amplify defaults to 0dB, your peaks are already hitting the 0dB maximum and the recording is probably clipped (or you were very lucky and your peaks just “touched” 0dB without trying to go over.)


  • Your ADC, DAC, and WAV files are hard-limited to 0dBFS. However, Audacity can go over 0dB internally/temporarily without clipping. So, you can boost the volume or boost the bass, etc., pushing the peaks over 0dB. Then you can safely use Amplify or Normalize to bring the levels down before exporting.

I’m not sure if you are familiar with youtubers such as markaplier or pewdiepie, but they are yelling and screaming in every video they put out. They have zero clipping or distortion, but I do from laughing. I don’t understand how they do that. Would you have an explanation on how they achieve this?

Ask them and then come back and tell us how they do it. It’s a forum, not a help desk.

You can without question get sound recording equipment that self-levels the sound as it is performed. Most cellphones and actual telephones try to do that. You can also record so the loudest possible sound doesn’t clip or overload and then compress and process the show in post production to bring up the low rather than reduce the loud. You don’t need anything fancy to do that. Just pay attention.

As DVDDoug above, this works a whole lot better if you’re recording in a quiet room with good microphone and recording electronics. A studio. There are two limits. If you get too loud, the digital system runs out of numbers and crashes (clipping). If you get too low, noise is going to get you. You can get noise from your microphone (ffffff) and you can get it from the room like that MetroBus driving by. They seriously limit the processing you can do and how crazy you can get on mic.

Many recording systems do loudness processing and it’s not up to you at all. I have a new recorder that if I scream into the microphone, it will just barely think about clipping…maybe. I have a good room for recording and the recorder doesn’t make very much noise, so that works out.

So it’s a juggling act.

I know what you’re doing, but It’s interesting to look at the AudioBook people. Their publishing specs are for peak, loudness and noise… and you have to pass all three in the same sound file or the automated tester will bounce you.

I do remember listening to the Pando Podcast and that one is solid yelling at each other. I doubt I could tell if the sound was noisy or not because they never shut up. Holes in the show are filled with music.

So there are production ways around this.


If it wasn’t clipping before you uploaded it to YouTube , you may not have left enough headroom

When you upload to YouTube you have to leave [ 3dB?] headroom , otherwise when YouTube transcodes your video into their in-house video format, clipping will be introduced.