How Audacity Compression works - help

Windows 10 Audacity version 3.0.5

I’ve read the Audacity instructions, I have done basic editing for years but I am a newbie to compression. I understand the concept of what it can do.

However in Audacity when I attempt to define the threshold input level, I cannot mentally map this level with the audio level shown in the waveform display (circled in red). To be specific, the threshold level in the compressor is set here at -10 db. Where does the -10 db map to in the audacity waveform shown by the red outline? The waveform there fits between -1 and + 1.

Thank you!

Montclair, NJ

Compression kicks-in (and the level is reduced) when the sound is above the threshold. When the sound is below the threshold it’s not changed.* The threshold is the “knee” in the compression curve.

(circled in red).

-10dB is about 0.3 or 30% on the linear scale. (I have the dB formulas in a spreadsheet. :wink: )


  • I’m not exactly sure what the Noise Floor does but I do understand the idea behind it… Compression is normally used with make-up gain to bring-up the overall loudness. That means quiet parts (including background noise) are increased. The Noise Floor setting is supposed to compensate for that.

See just above the “OK” button on the compressor, there is a checkbox, “Compress based on peaks”.
When that is selected (enabled), the threshold level maps to the peak level of the waveform.

Quick conversion from linear scale to dB:

Linear  dB
  0     -inf
  0.125 -18
  0.25  -12 
  0.5   -6
  1.0    0

(You can also view the track with dB:

When that checkbox is not selected, then the threshold level maps to the RMS (root mean square) level of the waveform. The RMS level is shown in Audacity as the pale blue region within the waveform.

Personally I find the compressor much easier to use with “Compress based on peaks” selected.

am a newbie to compression. I understand the concept of what it can do.

In general dynamic compression reduces the dynamic contrast by making the loud parts quieter and/or the quiet parts louder.

Most compressors reduce the loud parts and then make-up gain is frequently used to bring-up the overall-average loudness.

Limiting is a kind of fast compression, and again it’s often used with make-up gain to bring-up the overall loudness. If you just want to gain loudness, limiting is often better because it usually has instant attack. (The Audacity limiter actually has look-ahead.) There are also fewer settings to mess with so it’s usually easier to use than “regular” compression.

Automatic volume control (AKA automatic gain control or leveling) is a kind of slow compression.

Dynamic expansion is the opposite. It makes the loud parts louder or the quiet parts quieter. Expansion is rarely used in audio production except as a noise gate. A noise gate is downward expansion where quiet parts (hopefully where there is noise-only) are turned-down or silenced completely.

To get you going with the visualization thing, the blue waves are displayed in percent. 1.0 (100%) is 0dB. 50% is -6dB. The waves keep getting tinier and tinier until you can barely see them at about -24dB.

Many sound programs display in percent, but label it in dB. This is Cool Edit.

But, wait, I can hear you say, the bouncing sound meter keeps going to -60dB.

True, but most of the important stuff happens in that loudest 24dB and besides, it’s easy to program. You can force the blue waves to display in actual dB. It’s less handy than you think. Right-click in the percent column > dB.

I am a newbie to compression.

Do you have an actual job? The Audacity compressor can be a little rough to use because you have to accurately define the input levels and the expected reaction. What most people want is “push this button and the levels come out right.”

There’s two processes which can do that. Audiobook Mastering has tools that can put your voice volume exactly where you want no matter where you started, and Chris’s Compressor which does basically the same thing for music.

They both have rules.

Let us know.


Steve and DVDDoug and kozikowski, thank you for this so that I understand. I have no audio project before me at the moment but I wanted to practice on compression for the future. In the past I have manually compressed and expanded audio for videos and podcasts I have been given and thus I have spent much time working on them but they turned out quite good.