High Pass Filter: The First Step to Reduce Recording Noise
Are you using a High Pass Filter for eLearning Voiceovers? Do you know what that is and if so, how to use it effectively?
All of. your questions will be answered here! I was recently talking with an eLearning developer I’ve worked with for a long time and she told me that she has a lot of projects that she and other developers narrate themselves. She said that she uses the noise reduction in Adobe Audition or Audacity and they work great. But sometimes, there’s still a lot of noise in her eLearning recordings. I had to ask her:
“Do you run a High Pass Filter first?”
“A high pass…what’s that?”
So, I told her what I’m about to tell you now. A High Pass Filter (HPF for short) is a kind of wide-ranging EQ to apply over all of the low rumbling frequencies in your voiceover recording. Most of the noise that you hear in your recordings are down in the low frequencies. These are the sounds beneath the frequencies of the human voice: HVAC rumble, distant cars, a distant dishwasher on the other side of the house. These frequencies appear noticeably from 20-80Hz (Hertz). What if there was a magic sound eraser that makes all those frequencies go away and leave the frequencies of the human voice above 80Hz? There is! That’s the High Pass Filter! And you have it ready to use in both Audition and Audacity.
What is the High Pass Filter and How does it Work?
The picture above shows a High Pass Filter for eLearning voiceover (1) being applied to the frequencies of my voice, which are the squiggly white waveform lines. Across the top of the graph, we see the frequency starting at 0 on the left and going up past 10 Hz on the right. You’ll notice that with (6) I also engaged a +2 dB increase in EQ of frequencies at “10k” (10 kHz) to brighten up the high end, and to give some contrast for what I’m doing with the HPF.
If you look at the squiggly lines of my voice, you’ll notice it starts to be noticeable around 100 Hz and the first big bump is at 130 Hz. So we want to pull down all the un-needed rumble and noise below 80 Hz (Hertz). Note that the thicker white curved line of all EQ starts to pull down frequencies above 80 Hz. This is normal, unless you use a “brick wall” HPF, which doesn’t work very well on voice over and sounds unnatural, trust me. This is why we set the HPF to 80 Hz for voice. If you set it higher, to say 100 Hz or 120 Hz, that could be do-able for higher pitched voices, but you will be pulling down frequencies from up to 300 Hz, which can make the voice sound thin.
A common custom setting for the High Pass Filter is the “slope” or how steep it is. Going from gentle to steep, the HPF is usually available in steps down of -6 dB, -12 dB, -18 dB, and -24 dB per octave. Settings at -12 and -18 dB per octave tend to work the best, cutting out the most rumble without being too severe.
Noise Reduction After the HPF
So, now with the HPF engaged, we’re effectively pulling down, or “attenuating” all of the frequencies you see on the left in the purple area.
With that taken care of, you can then run your noise reduction program either as a second pass of processing or as a second part of your effects chain.
But here’s the great thing: that noise reduction effect won’t need to work so hard to get rid of noise as it used to do. You may find that now with an High Pass Filter for eLearning first in the effects chain, your noise reduction settings can be more subtle, which will lead to a more natural sounding eLearning voiceover.