Broken, Glitchy Audio


I recorded some voice, and when I went to play it back, I found out that it was completely unusable. The whole way through the recording is filled with an awful glitchy cracking noise, that no amount of noise removal can get rid of without making the actual voice incomprehensible. I’m using Audacity 2.1.3, and have never had issues before. I’m using Windows 10, and was recording from a USB microphone plugged directly into the port.

I am not able to redo this recording. Is there any way I can salvage it?
2019-02-05 16_11_08-Audacity.png

It’s probably [u]dropouts[/u], which are related to multitasking.

The 1st thing I’d recommend is upgrading to the latest version (2.3.0) which will show dropouts. Then, if that’s the cause you can work-on trying to prevent them.

Try to minimize any background operations. It may also help to increase your [u]Latency/Buffer Length[/u].

It gets complicated. There is a FREE online book about optimizing you computer for audio called [u]Glitch Free[/u].

…Your computer is multitasking even if you’re running only one application. The digital audio data flows-into a buffer (like a holding tank) at a smooth-constant rate. When the operating system gets around to it, the data is written to your hard drive in a quick burst. If the buffer isn’t read in time you get buffer overflow and a glitch.

When you play-back it works the opposite way. The playback buffer is filled in a quick burst and the data flows-out at a smooth-constant rate. If the buffer isn’t re-filled in time you get buffer underflow and a glitch.

Note that you don’t have to be utilizing a lot of CPU time. Some application, driver, or process, just has to “hog” the CPU for a few milliseconds too long and you get a glitch.

I don’t think that’s the problem. I think it might be worse than it looks in waveform. I’ve included a few seconds of the audio to demonstrate.

I think it might be worse than it looks in waveform

It looks pretty bad in the waveform. The blue waves only show the top 25dB or so of the show sound out of about 96dB. In English, anything that appears on the timeline is going to be a significant percentage of the show.

The show is trash, but I’m more interested in the ratty voice. How was this shot and with what? Are you sitting in the back of an auditorium? It sounds like a bad phone recording past the obvious interference.

And No, I don’t think those are dropouts. Are you standing on a manufacturing floor or other heavy industry location? There are two different pitch buzzes in there. A different one starts up at about 2-1/2 seconds.


Oh, wait. I’m asleep. Cellphone. That’s classic cellphone negotiation and connection management. Were you holding a cellphone close to a consumer-grade USB microphone? It doesn’t matter if it says “Studio Professional Broadcast Microphone” on the box. If it cost much under about $100, it’s a home microphone and not shielded against radio interference.

Don’t record that way any more.

I would be turning my cellphone off. If you’re in an audience, it can be anybody near you, too.


Doesn’t have to be direct radiation into the microphone. That USB cable makes a dandy antenna. Are you using a standard USB cable about 6’ (2M) long? If you have a standard length cheap cable or an extended cable, those can more easily pick up trash from the air.

A lot of this is guesswork and we can go further when you post how you shot this.


Yes, I was holding a cell phone next to it. I have done it before, but the difference this time is that I was using cellular data instead of WiFi, which I guess caused the interference… I was recording an interview with someone over Skype. Lesson learned, I guess…

Skype has a new service where they will record your interview for you and provide a soundtrack to download.

Google your brains out.


I’m assuming there’s no way to silence the interference?

which I guess caused the interference

You don’t have to guess at it. Try it different ways as a test and see what happens. If you’re doing this interview stuff for real, you should know all the ways your recording system works and all the ways it fails. Start recording and announce as you go.

“This is my cellphone on cell service three feet away.”

You can also make Audacity feed your recording audio back to your headphones as you try different variations in real time. The sound will have an echo from computer delays, but you’re not recording the Boston Symphony Orchestra. You just want to know how close you can get a cellphone before your system drops dead. Try it on WiFi, too, to get the difference.

Once you know how to cause the damage whenever you want, try a shorter USB cable or just a different one. Some of the cheaper USB cables don’t have good shielding under the rubber cover.

Sometimes you can help stress situations by wrapping the microphone in aluminum foil except for the front. Check the sound doesn’t get screwed up if you do that, but check it ahead of time. Not by losing the next interview.


Sadly, no. The damage is far too severe.

It should be fairly straightforward to test this theory. See if you can reproduce the problem AND test if you can avoid the problem (by moving the cellphone away from the mic).

I’m assuming there’s no way to silence the interference?

That’s correct. It’s too loud and it has a lot of the same tones as a voice which is why Noise Reduction failed.

You can make it sound like a bad cellphone.

That’s taking the profile from that first clean segment. Drag-select that.

Screen Shot 2019-02-05 at 19.00.01.png

Effect > Noise Reduction > Profile > OK.

Then select the whole clip by clicking just right of the up arrow (on the left) and Effect > Noise Reduction > 18, 6, 6 > OK
You have to do custom noise reduction every time the phone changes gears, and to do that, he has to stop talking long enough to get the Noise Reduction Profile. I’m guessing he doesn’t.