Noise Reduction/Removal

It’s likely I’m going to give you corrections that only work when applied to the original read. It’a a very newbie mistake to apply corrections to a performance and then save the new file on top of the old file—reuse the old filenames. This destroys the original read making it impossible to “go back” and correct the work.

Just from a quick listen to your submission, I would also want to know about your studio. Who made the microphone and how is it connected? Pretend I want to buy what you have.


There is a VERY loud tone at 90Hz . That can be quickly removed with a notch filter …
notch filter @ 90Hz Q=10.png

The latest version of Audacity has a notch-filter in the effects list . If your version of Audacity doesn’t you can get one here.

De-clicking is a far more time-consuming and computationally-intensive task than removing constant hum with a simple notch-filter.
I used the [beta?] Paul-L de-click plugin in Audacity , which you can get here …
However that plugin can easily take 5x the playback-time to declick audio. I’ve only used the Paul-L de-click plugin on very a few minutes of audio in one go , others have complained a crash occurred when they tried to process hours of audio in one go.

Trebor, THANK YOU!! That worked beautifully. Any idea where that 90Hz tone might have come from? Recording conditions were identical to other recordings I’ve made that had no background noise.

Koz, I am in fact a newbie to Audacity, but I do save all of my raw recordings :slight_smile: Attached is a sample from the original read. Again, I’m quite satisfied with the fix Trebor provided, but I would also love any additional corrections. I’m using Windows 7 Home Premium, Audacity 2.1.0, .exe installer.

My “studio” at the moment is my walk-in closet. I have a small table set up, on which are sat the PylePro PSMRS11 Microphone Isolation Shield and my laptop. I set two pieces of Pyle PS11612 soundproofing panels on top of the shield, as well as set my laptop on one.,-Recording-Foam-Wall-Tiles,-12-x-12-Squares,-1-Thick-(12-Pieces), and I’m working with the Focusrite Scarlett Studio pack


90Hz is such a weird tone.

I think I got it. The stock Focusrite Scarlett Studio pack (nothing wrong with anything you’re doing, by the way) doesn’t have a shock mount. That’s typically a rubbery, spidery, jiggly thing that holds the microphone. I made one out of plumbing parts and US Postal Service rubber bands. Used it that way a couple of times, too.

The idea is you can have floor or desk vibration that never makes it up to the microphone because of the sloppiness of the rubber bands.

Note in the illustration for the sound shield, they show a microphone with a shock mount (attached).

I bet you’re listening to air conditioning pumps or other low pitch ventilation system noises coming up through your desk. As a test, you can disconnect the microphone from the base and just hand-hold it in about the right place. Don’t let it touch anything but your hands. Make a test recording. Yes, I’m clear you’re going to get tiny hand-holding noises, but I bet the pump sounds go away.

Does that microphone have the option of a shock mount?

Let us know.

Screen Shot 2015-08-17 at 13.20.11.png

A completely side issue is recording volume. As a fuzzy generality, the Audacity recording meters should jiggle around -6, the yellow zone.

Your voice volumes are WAY low. Again with fuzzy arithmetic, I have to boost your voice 20dB to make ACX AudioBook volume. In English, sound doubles every 6, so I have to double your volume three times to make ACX volume.

You should avoid running the sound meters all the way up (the red zone) and The Scarlett Solo’s knobs will flash red if you get too loud.


The Paul-L De-Clicker plugin* does do a good job , but takes its time. With the settings shown below the processing time is approximately three times the play back time …

The effect settings used on the audio file also attached to this post.png
[ * Audacity does not come with Paul-L’s deClicker installed , you can get a copy of it from here : the “Declicker.ny” file at the bottom of the post . Put a copy of that “Declicker.ny” file into Audacity’s plugins folder , then it can be included in Audacity’s “Effect” list via “manage” at the top of the effect list ].

And you won’t need some of those corrections if it turns out there’s something you can do to the physical studio to get rid of that 90Hz tone. That will speed things up.


Thanks, guys!!

My friend said he can rig me a shock mount, much like the homemade one you described, Koz. I’m sure you’re right that something is coming up through the table. Could be the A/C.

So Koz, in order to get my super quiet recordings up to ACX standards, I would run the Amplify effect approximately 3 times? That’s one thing that I haven’t been able to figure out – how do I know when the overall volume is within the submission requirements (where on the recording do I look to find that info)? I posted that question in another thread (technically that + how do I ensure the volume of all my chapters is consistent), but no one has bitten yet, and I’m trying to take advantage of your expertise while I have your attention :slight_smile: Your help is greatly appreciated.

And I wonder why my volume is so low in the first place. I record fairly close to the mic (3-4 inches). Is this as simple as turning up the gain when I record?


Shoot. OK, so in the mastering process, should I amplify first until I get to a submission-ready level, then perform the other steps? I did the process in the ACX -recommended order (Hi and Lo Pass filters, Normalize peaks, Equalize, Limit), but it was too quiet, so last, I amplified it. Well now the peaks are way above -6dB… :imp: Where did I go wrong?

Also, everyone who’s been contributing to this thread, I’m going to need your home addresses so I can send you a beer or something - I would truly be lost without your help! :wink:

Trebor, I just ran into a similar problem that you solved so nicely with the notch filter a few months ago – hoping you (or anyone!) can diagnose this one as well. Same situation - weird tone, noise reduction doesn’t fix it. I just recorded my first dual-narrator project, and we did it in the other narrator’s home studio. This same tone pops up every once in a while - I’m wondering if it happens when one of us accidentally steps on the headphones cord or bumps the mic in a certain way?? I’ve tried messing around with various frequencies to see if I can find the culprit, but to no avail. I’m using Windows 7 Home Premium, Audacity 2.1.0, .exe installer. Thank you!!

There is a fairly constant tone on that at 44Hz , but I can only just hear it , it’s visible on the spectrogram …
annotated spectrogram of ''UnidentifiedTone_CG''.png
You could remove the 44Hz tone with a notch-filter , but I don’t think that tone is the intermittent noise you are hearing.

The conspicuous intermittent problem is excessive-sibilance : when the speaker says “realiZe” “childiSH.
The cure for excessive-sibilance is a de-esser plugin* , e.g. Paul-L’s De-esser.

There are also a few clicks, e.g in “they’re” , Paul-L’s De-clicker will get rid of those …

[ * Other free de-esser plugins are available which work in Audacity on Windows, e.g. Sibalance [sic] , SpitFish ]

Thanks so much!! Awesome, awesome, awesome. I will get those two plug-ins today and keep you posted as to how it goes. I’m a “clicker,” almost all the freaking time – I haven’t yet figured out how not to be so clicky. Most days I am; some days I sound crystal clear. Have high hopes for that De-Clicker!!

Turns out the clear days are when you’re slightly buzzed on Chablis. Even worse would be the invitation from the local school for a lecture about your announcing secrets.

Warm lemon and honey?

No, more like Domaine Louis Moreau in a large glass.


LOL thanks Koz. But also, oh no! There is definitely a tone that keeps popping up - a few times now it’s responded to the 44Hz notch filter, but there are some long sections where it’s not (I’ve played around with all the Hertzes [that’s correct; I promise :stuck_out_tongue:] from 40-60 and nothing is doing the trick. Help please!! You guys are the best.

ARGHHHHH so frustrating because it seems to go in and out… hoping there’s a quick(ish) fix?? The attached sample is longer and (along with some fun mouth noise - you’re welcome) you can hear it getting more and less intense…

On that one the main culprit is 79Hz , (& harmonic at 158Hz ) …
Unidentified-Tone_CG wav  (182 04 KiB).png
The pitch is changing slightly with time : maybe a fan ?, or other electric-motor ,
( So it’s not mains hum, whose pitch is very constant ).
Notch-filters are only 100% effective when the pitch (frequency) is constant.

Your [female] voice doesn’t contain frequencies below 150Hz , so using the equalizer you could remove all content below 150Hz , which will remove the deep parts of the humm noise …
equalization to remove all frequencies below 150Hz.png
In addition you’d have to apply a notch-filter to remove the 158Hz tone.

Uh-oh – when I apply those two notch filters (79 and 158Hz), it almost entirely eliminates the hum, but it also eliminates the lovely resonance in my co-narrator’s voice (cf attached clip)… Any ideas?? Maybe the undiscerning audio book listener won’t notice the hum?! I’m freaking out a teensy bit, as re-recording the entire project really is not an option, and this tone shows up quite frequently. Is it possible there’s a separate fix I can apply to the hum when I’m speaking vs. when he’s speaking? Crazy tedious, but again, better than re-recording. Would it be helpful for me to upload (or PM) a longer sample - one that contains the hum throughout, plus both of our voices? Thanks so much.

I can’t hear any hum on the “before” on that one, when the man is speaking.
[ The equalization suggestion in my previous post could only work for a female voice ].

If you split your recording into two synchronised tracks : one with everything above 200Hz , the other with everything below 200Hz, then apply a dynamic-range-expander, ( e.g FLOORFISH ) , to the sub-200Hz track ,
that will selectively attenuate the bass [hum] when the man isn’t speaking . Then recombine the two tracks.

Hmmm. Yes… if I just split the track… :confused: How would I go about doing that?

BUT if you couldn’t hear the hum… and I also had someone else listen to one of the chapters where the hum is definitely there (I can see it on the sound meter and feel it vibrating my ear drum), and he didn’t notice anything wrong… would it be a fair assumption that your average audio book listener won’t be freaking out about it in the way that I am?

Maybe the undiscerning audio book listener won’t notice the hum?! …

They may not. But ACX will. Or the Robot will get there before human Quality Control.

We simulated what the robot is going to do with Will’s compilation testing tool called ACX Check (if I haven’t mentioned that already). It’s a plugin that appears under Analyze.

That may be the older one. The newer one is prettier, but says the same things. I gotta fix that one of these days.

ACX has a very simple weed-out process that catches ratty submissions. When you’re speaking, your peak sound values have to stay below -3dB, your volume or loudness has to be between -18dB and -23dB, and when you stop talking, what’s left (the background noise or Room Tone) has to settle below -60. If you do a very good quality job of announcing, you can crank out one of these compliant clips without a big deal, or your clip can be made complaint with a simple volume change. I, my very self, have done this, so I know it can be done. It’s not fantasy and it also happens to be broadcast standard.

Then it’s on to the human Quality Control. If you got past the Robot by applying legions of filters, effects, patches, modifications, and changes, you will most likely fail QC with “Excessive Processing” (he said, gazing up at the above thread’s list of notches and filters).

Without knowing any more about your setup than the amount of corrections it appears to require, I would think about changing it.

You can achieve the -3 thing with Normalize. I use -3.2 instead of -3 for safety. Conversion to MP3 later can mess with this setting.

– Select the whole clip or show by clicking just above MUTE.
– Effect > Normalize: [X]Remove DC, [X]Normalize to -3.2 > OK

Then Analyze > ACX Check and see what it says.