Degradation due to sound removal

When I use the sound removal tool my entire sound track is degraded. How can I fix this problem?

Noise Reduction works by removing sound from the performance. It removes the trash that you don’t want in addition to a little sound that you do. It’s unavoidable. It pokes holes in the show. There will always be some degradation of the show when you use Noise Reduction.

Noise Reduction is always a compromise between the desirable and undesirable sound that’s why the tool has the adjusters and sliders. Keep suppressing noise until the quality of the show suffers. It’s very rare for one simple setting to work for all shows and selecting the profile, or “Just Noise” is very important.

That’s also why it’s of paramount importance to shoot the best quality show you can and not need noise removal at all. Noise is one way you can kill a show – it’s one of the four horsemen.

You should be using Audacity 2. The noise reduction in the newer Audacity versions works better than the older ones.


Thank you for your quick response. I don’t always have the luxury of setting a scene or control over a particular microphone, So I have to do everything postproduction.

Can you tell me which way to pull each of the sliders to get each effect?

I don’t always have the luxury of setting a scene or control over a particular microphone

That never happens to me…

We should be very clear that the Profile step be correct. Select noise by itself without any performance in the selection. Without that, the tool is severely compromised.

Noise Reduction is the amount of correction that’s applied. If that’s at zero, it doesn’t matter what the other sliders do. Preset around 6 to 12.

Frequency Smoothing is the amount of effort Audacity goes to avoid people talking or other valuable work. This prevents martian/robot voices. At zero, it’s unrestrained and applies reduction to everything. At maximum, it will give very noisy voices and other work and only reduces noise in the spaces between voices. Tune for clear voices and performance.

Attack/Decay gives the lead and lag for the correction. Sometimes, it’s best to start the correction a little early and stop it a little late. When you walk through a door, the door isn’t exactly the same size you are. It’s a little larger.

Sensitivity I intentionally left that one for last. I have no idea what it does. I don’t see a pre-baked tutorial, either. I’ve never messed with it.


The manual page is here:

It adjusts the level of the noise profile.
Increasing the sensitivity removes more “stuff” and decreasing the sensitivity removes less “stuff”.
Ideally the “stuff” should be just noise but in practice it is noise plus anything else in the recording that has a similar frequency and amplitude as the noise profile.

Normally the Noise Removal effect will attack the noise that occurs “between” other sounds. Increasing the sensitivity setting can help to reduce noise that occurs “within” other sounds, though it may also cause more damage to the other sounds.

Usually the sensitivity should be left at the default value of zero.