Noise Floor

I’m using Windows 10, Audacity version 2.2.1. When I go to Effect/Noise Reduction/Get Noise Profile, nothing happens. Should there be a pop-up to tell me what my noise floor is?

That’s how it is : no notification that a noise-profile has been captured.
Audacity will remember that noise-profile until you get another noise profile, (or until you close all Audacity windows).

Noise Reduction works in two passes. Select something from your performance that’s pure noise. The profile step allows Audacity to “sniff” the noise to know what to do. Select the whole show and run Noise Reduction again with the settings you want. This time Noise Reduction does the work.

There are some restrictions. If the show is so noisy you can’t find the “pure noise” parts, then that may be the kiss of death. Record it again.

This only works on constant noises. We can’t do a thing for moving or changing noises. That metrobus going by is now a permanent performer in your show.


I’m missing something here. There’s a window below that “Get Noise Profile” button.I selected a silent section of a published audio book and the Noise Reduction Db is 12. Then I selected the silent parts of an audio sample that I sent you guys that had a bad noise floor problem, and that number is also 12. Nothing changes, nothing happens when I push that button. I need to see what you guys are seeing so I can fix it.

Sorry, I made an error. The second file I tried to get info on, I sent to the ACX team, not you. They said it had a bad noise floor problem.

That’s the amount of noise reduction… When it “finds” noise it will be reduced by 12dB. If the number is higher the noise will be reduced more but I assume there’s a higher chance of audible artifacts.

After selecting the “silence” and getting a noise profile, you then have to select all the audio you want the noise-reduction effect to be applied to …
'Noise Reduction' is a 2 step process.gif

The Profile step has no graphics. You click Get Profile and everything vanishes. It doesn’t tell you a thing.

It’s the second pass that allows you to set how you want Noise Reduction to work.

Don’t tell ACX you’re using Noise Reduction. They hate it (mostly because everybody uses it wrong).

Do you want us to drop in to the mastering tennis match? I can tell you the specific steps to pass ACX technical conformance.

Post one of these. It’s pretty simple. The hardest part is not moving around or gasping for two seconds.


Thanks Koz! I’ll try this today. :slight_smile:

Hi again,
Dropping in to the mastering tennis match, I don’t know where the forum text entry windows: “Upload attachment” tab is. Do I save the exported file then go somewhere else? Thanks, Carol

Upload attachment tab.gif

There are some cautions with that. You can’t upload massive productions. It will only allow you (for example) 20 seconds of mono WAV or 10 seconds of stereo WAV. Uncompressed WAV (Microsoft) 16-bit is recommended if you want us to evaluate anything because the world’s favorite format MP3 has built-in distortion and we can’t tell which damage is yours.

There are also restrictions for file types.


Hi, Can you tell me what the floor noise is and give me any other useful info that would help me meet ACX requirements. This is a Yeti mic. I have another large diaphragm mic and need to know which one will work best. Thanks!

Yes, but.

It’s a little messy, but fixable in the Audacity tools. Repaired clip attached.

First three sentences and declaration 2/3 down.
Screen Shot 2018-05-24 at 13.08.57.png
It’s not noisy. You are.

[loud lip tick] She walked into the bedroom…

…the girl [loud gasp] She walked into…

…behind her [loud gasp] Five minutes later…

I doubt Quality Control is going to pass that for purely theatrical reasons. You can fix some human noises in post production editing, but it might take weeks for a book. There are some tools that can help with mouth smacks, but I think your restricted airway could be a problem.

I like your free-swing delivery. I had a friend from North Carolina who could do that. “OK, you just sit right here, I’ll get you a coffee…and a microphone.”

I think you’ll have to surgically removes all those gasps…by hand.


It can be done. It’s intensely painful editing.


Removing them is not just cutting them out or silencing them. One throws off the story timing and the other puts odd holes in the room sound behind you.

I did it by drag-selecting a gasp and then looking at how long the time is (.63 seconds). Then I copied a 0.63 second chunk of normal room sound and pasted it over the gasp. Repeat many times. I also cut out the lip smack at the beginning. That’s a straight drag-select and cut.


As a last obsessive act, I removed a bit of silence from between each phrase to “tighten it up” a little.


You are awesome!!! Thanks! I’m going to send you another one with a different mic, paying attention to all mouth sounds and breaths. This helps SO much!

Hi Koz,
This upload was recorded on a large diaphragm mic. The upload I sent you before was done on a Yeti. Is there a difference that you can discern? Which one is better? I have learned how to edit out loud breathes, fill in or take out spaces. Next learning curve will be to cut and paste. :slight_smile: Really appreciate your help. This should be the last question.

Niagara Falls

I like this one for voice tone quality. I think your story-telling swing is better, too.


It has a very serious power hum problem.

mmmmmmmmThat nightmmmmmmdressed alike in dark green, bell bottom pantsuitesmmmmmmmmmmm.

What’s the microphone and how do you have it connected?