A compressor is a device for automatically controlling the level of an audio signal. Specifically, a compressor reduces the volume of an audio signal when the volume exceeds a threshold set by the user. A limiter is a specialised form of compressor that effectively limits the input signal to the threshold level.

The Controls

Typical controls you will find on a compressor are:

1. Threshold

2. Ratio

3. Attack

4. Hold

5. Release

6. Gain

7. Bypass



The threshold control is used to set the threshold level. This level is the level, in dB, above which compression will be applied to the input signal.



The amount that the signal is attenuated by is dependent on the selected compression ratio. For example, if the compression ratio is 4:1 and the input signal exceeds the threshold level by 4 dB the output signal will only exceed the threshold by 1 dB. When the signal falls below the threshold no compression will be applied. If the compression ratio is set to 20:1 or more the compressor is being used as what is known as a limiter. This is because the signal is effectively being limited to the threshold level. Most compressors have a broad enough range of compression ratios that they can operate both as compressors and limiters.



Attack, Hold and Release

To make a signal sound more natural at the moment of applying or removing compression, i.e. the signal crosses the threshold level, most compressors provide the ability for the user to adjust the attack and release times. In some compressors these controls are automated. Using the attack and release times, the application of the effect can be made less blatant and the transition smoother.

Attack time: How long after the signal exceeds the threshold level that compression will be applied.

Release: How long after the signal drops below the threshold level before the gain/attenuation returns to uncompressed levels.

Hold: A minimum time before the release phase of compression will be entered.

If Attack and especially Release times are very short, the resulting rapid changes in gain create an effect called either "Pumping" or "Breathing". This simply means that the compression is an audible effect, rather than a subtler enhancement. If the attack and release times are too short bass frequencies can be distorted as the gain rapidly changes. A hold time of about 50ms will prevent this happening to audible frequencies.



Primarily, compressors reduce gain. To address this most compressors have an output, or "make up", gain control. This basically allows the peak level of the input signal to be maintained. The overall process of compressing the signal above the threshold and then boosting the overall signal amplitude reduces the difference in amplitude between the loudest and softest sound.


Compression Details

The Knee

A basic compressor does not affect the signal until it has exceeded the threshold volume. At this point, delayed only by the attack time, it fully applies the compression selected by the user. This type of compression is known as Hard Knee compression. This may be sufficient for overall level control but the effect becomes too obvious on more heavily compressed signals at the front of the mix.

Click on the diagram to see a larger version of the image.


Hard Knee Compression

A smoother compression can be obtained using Soft Knee compression. This applies an increasing level of compression gradually to the input signal as the signal level approaches the threshold level. This ensures the transition from uncompressed signal to compressed signal is far less noticeable.

Click on the diagram to see a larger version of the image.


Soft Knee Compression


When would you use Hard Knee?

There are a few cases where you might use Hard Knee compression:

1. When you are using compression deliberately as an audible effect. Normally you would use a fairly heavy compression if you are using it as a noticeable effect.

2. A small amount of intentional pumping can give an overall impression of loudness. A Soft Knee compressor detracts from an obvious pumping effect, so a Hard Knee compressor is preferred.

3. If you are using high compression ratios Hard Knee compression will give a firmer gain control. Different Soft Knee compressors have different characteristics. Some compressors have a small knee that increases the compression over only a few dBs, where as other compressors have a relatively large knee by beginning to attenuate the input signal at 20 to 30 dB below the threshold.


Side Chain

This is the part of the compressor that measures the level of the signal to be compressed. Different compressors have their own characteristic sound partly due to the differing side chain circuitry, for example, different responses to changes in input level or frequency sensitivity due to filtering. Filtering is often applied to the side chain signal in order to compensate for the way humans perceive different pitches, of the same amplitude, as having a different loudness. If filtering is not applied by the compressor in the side chain circuitry the compressor may appear to respond differently to some frequencies.

Many compressors provide access to the side chain circuitry input, an insert point often called the Key. This input allows other processors to be inserted into the side chain or for the side chain to be triggered by a completely different input than the signal being compressed. A common application of this could be the insertion of an equaliser in the side chain to make it sensitive to the high frequency sibilance of a vocal track. This is effectively De-essing. Another common application of side chain circuitry is to trigger the compressor from a separate source. This is commonly used by a DJs etc who wants to turn the music down whenever the speak to the audience and when they stop talking the volume of the music returns to normal. This effect is known as Ducking.



Some compressors have a built in gate or expander to mute the output of the compressor whenever the input signal drops below a gate threshold. This is because a simple compressor cannot differentiate between low level signal and unwanted low level noise, such as the quiet sections between phrases in a vocal track.

Discuss this article in our Music Forum.


John Moxey

Songstuff Site Crew


About Songstuff

Logo-dark-bg-square-170Songstuff was launch in September 2000 and has grown into an all round resource for musicians, attracting interest from musicians of all experience and skill levels. The Songstuff Songwriting and Music Community has grown into an essential, dynamic networking resource, where members exchange ideas and collaborate on common projects. Great thanks and appreciation are owed to the moderation team for helping to grow the community into the active and creative place it has become.

Site Crew conduct draw on their experience and contacts to perform interviews, and write quality articles on a variety of subjects. In addition the Songstuff Community members regularly contribute articles and Songstuff has many regular contributors from across the field of music.


Our Crew

Songstuff Site Crew are highly experienced and cover a broad range of music industry roles including label owners, music educators, professional musicans, songwriters, band managers and other music industry professionals.

This article has been written by one of the Songstuff Site Crew.

Songstuff Home Page

Contact Songstuff

People compose for many reasons: to become immortal; because the pianoforte happens to be open; to become a millionaire; because of the praise of friends; because they have looked into a pair of beautiful eyes;”