I use the noise removal filter all the time. It removes the background noise whenever they’re not speaking, but if there’s a lot of background noise it always sounds like a wave is flowing over the microphone whenever they start talking. Is there a way to remove the white noise from their voice as well as from the background?
I’m working on a video involving two people from two different locations. My audio was recorded with a Snowball microphone and I ran my usual compression and noise removal to even out the audio and get what little bit of noise there was out. But the second person’s audio has a lot of white noise that needs to be gone to avoid the wave sound effect and for us to have matching audio.
I know he could just rerecord his lines, but I’d rather not make more work for him.
Noise Removal can work well when there is constant low level noise, but if the noise level is high then that is a much harder problem. Getting the “best” results is all about compromise. For it to not sound weird (metallic, bubbly, watery, swooshy, robotic) you will probably have to settle for leaving some noise and just “reducing” (rather than “removing”) the noise.
Try pushing the “Sensitivity” control up to about 8 and lower the “Noise Reduction” setting. Increase the “Frequency smoothing” a bit (try around 500 Hz) and reduce the “Attack/decay” to 0. You will find that these settings will tend to lower the overall volume level, so you may need to apply a little “Amplification” after running Noise Removal. Experimentation is vital, but don’t expect great results - there’s good reason to use a good microphone and a quiet recording room
Hiss and “white noise” is almost impossible to manage because of the way Noise Removal works. You let it sniff pure noise and it generates custom filters to remove that noise from the show. Piece of cake, right? Hiss and White Noise contain all frequencies. So you are asking Noise Removal to remove all sound from the show.
OK, maybe not a piece of cake.
The only thing you have left is the trick of stopping the filter from working during voices. If the noise is bad enough, this gives you hissy voices over velvety silence. You are urged strongly not to shoot for perfection. It’s tempting to crank noise reduction up as far as it will go, but that’s a recipe for honky, bubbling sound. Best to sneak the correction in and stop when the valuable show sound starts to get damaged, whether or not the background is silent.
Noise Reduction is a violin, not a switch, and yes, it pays enormously to shoot sound perfectly during the live performance.
Snowballs work OK for desk announcing and other presenter jobs, but I wouldn’t pick one of those for video work. the best you can do is cardioid – receiving sound from the sides and front (a little sloppy) and very little from the back. You can reduce the sensitivity with the -10 switch, but you can’t ever increase it. That gives you the hiss if you have a low volume performance. That’s also one of the shortcomings of a USB microphone.
Shotguns are good. Very sharply directional with good volume and sensitivity – and only from the front. This is an extreme, crazy example from Hollywood, but you get the idea.
You would assume the sound might be compromised with this technique, but This American Life uses a shotgun for all its NPR face-to-face first person interviews – and not one of the crazy expensive ones, either.
I know people who do this. I need to hit them up for a picture.
