Often the most misunderstood piece of equipment is a Compressor. With such a wide variety of options to choose from, it's time we explain what a compressor is, how it works, and how to use it.
You've heard about them, read about them, they're hard to ignore even if you tried. Whether as software plugins, or outboard hardware, Compressors have been a staple in audio production for many years. Their use in recording, mixing, music production, TV and post production, as well as live performances, makes them one of the most used tool in an audio engineers toolbox. There are more brands and types of compressors than you can shake a stick at. People will proclaim their love and affection for certain compressors, myself included. To truly understand why compressors are used as much as they are, we have to understand how they work and how they are used.
What is a Compressor?
Simply put, a Compressor is an audio processor which reduces the Dynamic Range of any audio signal. (Dynamic Range is the distance between the softest sounds and loudest sounds) It appears the loud parts get softer, as the soft parts get louder. By reducing the distance between the softest parts and the loudest parts, the audio signal plays back at a more even, balanced level. During the process of recording, certain parts will be louder than other parts which is a problem for playback as you have to constantly "chase the volume". If you reduce the Dynamic Range (distance between softest and loudest) and Compress them closer together, the audio signal doesn't have as much distance between soft and loud. In the example below, the top image is the recorded audio with no compression, while the bottom picture is the same audio with compression. Notice how the distance between the softer parts and louder parts is reduced with compression applied.
In the original audio signal, there is a significat distance between the loudest peaks and the softer sound. As the compressor is applied, the peaks are reduced more and more, which creates a balance between soft and loud. If you keep on compressing the audio signal you can "squash" the sound, possibly making it sound lifeless.
The picture above is the same audio signal with a huge amount of compression applied. You can see that almost all the peaks are gone, with only a few poking out for brief moments. With this example being a vocal track, we have effectively made the softest part of the vocal equally as loud as everything else. This could also mean the breaths between words, background noise, and room ambience, would be just as loud as the words themselves. The hardest part of using a compressor is figuring out the balance between what the technology is capable of doing, when to use it, and how much to use. It's like walking a tight-rope.
How Does It Work
Before you let a compressor run around squashing everything in sight, it's a good idea to understand how it goes about processing audio signals. There are several controls that allow you to dictate how the compressor reacts to, and impacts audio. The main setting that triggers everything else is the Threshold setting.
Threshold
The threshold setting tells the compressor when to start compressing, based on whether or not the audio signal is loud enough to pass the threshold. Any audio that does not pass the threshold setting will not be affected by the compressor.
Threshold setting tells the compressor when to kick-in and start working. Audio that is not loud enough to pass the threshold will not be affected.
Several modern plugin compressors show you the threshold setting as a graph display like the one below. The vertical line is the threshold setting, the dot represents the audio signal.
Ratio
Once the audio signal passes the Threshold, the compressor needs to be told how much reduction will be applied. This variable setting is known as Ratio, and can be changed to allow you to work with different types of material at different dynamic levels. There are a wide range of Ratio setting depending on which brand, and style of compressor you are using.
The Ratio is usually represented as a numerical figure such as 1:1, 2:1, 4:1, 10:1, etc... which are represented in decibel values. The Ratio of 1:1 is the lowest setting, with no attenuation of the audio signal, basically this means no compression will be applied. A ratio of 2:1 will give you 1/2 the output level, for example: if an audio signal is 8db above the threshold, with a ratio of 2:1, will give you a final output of 4db. See picture below.
The greater the compression ratio, the more "squashed" a sound can appear. However, this is always dependent on the Threshold setting. If you have a high threshold, where only a few notes cross it, you might not hear too much of it. If the body of the sound sits well below the threshold, it might seem the compressor is not doing anything at all, even with a large Ratio setting. If you have a very low Threshold setting, where the main body of the sound always crosses it, the compression will be much more noticeable, which is where the "squashed" aspect comes in. The hardest part of using a compressor is knowing how far to push things without making them sound un-natural.
Attack and Release
Most modern compressors give you the ability to adjust how fast or slow the compressor reacts to the audio signal. Once you've set your Threshold, the Attack time tells the compressor how fast to react to the incoming signal in order to bring it to full compression. 
Sounds that have very fast transients, such as drums, might require a faster Attack time so the compressor may quickly grab and compress the sound.
The Release time is literally the opposite of Attack, it tells the compressor how fast to "let go" of the sound once it falls below the Threshold setting.
Knee
The vast majority of modern compressors will also have a Knee setting, which tells it how to transition from the non-compressed to compressed states. Is the transition slow and subtle, or is it sudden.
The Knee setting is most likely harder to hear as it is a very dependent on the material being played. A quick, short, sound will be harder to hear a difference in Knee setting (drums), as opposed to a long, sustained, sound that ramps up in volume (piano).
Gain
As you start using the compressor you'll notice that as you bring down the Threshold setting and turn up the Ratio value, the overall signal will become softer. To compensate for this, you can turn up the Output Gain (sometimes called Makeup Gain) of the compressor to "make up" for the loss in signal. 
Most compressors have a visual meter display which show you the amount of "Gain Reduction" (GR) you are experiencing in the audio signal. You can use this value as a baseline for how much to turn up your Gain setting. For example: if you see -6db Gain Reduction, you can turn up the Output Gain of the compressor by +6db.
Input meter, Output meter, Gain Reduction Meter
Types Of Compressors
There are a few main types of compressors you'll come across in the hardware world. Most software plugin compressors is designed to be an emulation of hardware, however, plugins can have several variation not possible in the hardware world. Other than plugins, the main types of compressors are:
Tube Compressor
This is the oldest compressor technology. Before there were circuit boards, there were vacuum tubes. Tube technology exhibits a certain characteristic due to how to tubes themselves react to signal. They tend to be slower with attach and release times, and often "color" the sound a lot more due to the harmonic distortion inherent in vacuum tubes. Examples of Tube Compressors include: Summit Audio TLA-100, D.W. Fearn VT-7, Undertone Audio UnFairchild, and many more. One thing to note about tube compressors, they tend to be quite expensive.
Summit Audio TLA-100A Tube Compressor
Optical Compressor
Many people tend to love Optical Compressors such as the Universal Audio LA-2A, especially on things such as guitars and bass. The technology is quite unique, as it uses a light element and an optical cell, which tend to be very smooth sounding. As the audio signal becomes louder, the light element emits more light, which causes the optical cell to reduce the output of the audio. Just as light on a dimmer switch, the movement from soft to bright light is very smooth and gradual. Example of an optical compressor include: Universal Audio LA-2A, Pendulum Audio OCL-2, Buzz Audio SOC-2, and many more.
Universal Audio LA-2A
FET Compressor
Due to the use of a "Field Effect Transistor", these types of compressors are very fast and clean. They do not add a lot of color like Tube compressors, and they tend to be much quicker as well. Generally, the slowest attack time on a FET compressor is faster than the fastest attack time on a Tube compressor. Examples of FET Compressors include: Universal Audio 1176LN, Daking FET Compressor II, Drawmer 1978, and many more.
Universal Audio 1176LN
VCA Compressor
While other compressors might seem flashier, at the heart of many mixes lies a solid-state VCA Compressor. VCA stand for "Voltage Controlled Amplifier" and are some of the most widely used of all compressors. Their circuits allow for a wide range of options, from very fast attack/release times, to very slow times, as if it were a tube compressor. Due to their versatility, you will even find VCA compressors on the Main Mix of a console, such as the SSL, Neve, etc... They are the "go-to" compressor for many people as they can be used across a wide range of material such as drums, vocals, guitars, as well as the entire mix.
Solid State Logic VCA Mix Bus Compressor
Practical Application
Let the fun begin! This is where many of us have found it takes some time and practice to get the hang of using compression day to day. First of all, you have to figure out if the sounds you're dealing with even need compression. Often times things such as samples have compression applied already. Or, the sound might be a light string part layered in the background, that might not need a lot of compression to sound correct in a mix. Also, there is No wrong way to use any equipment in the studio. The best advice I ever received was "Just turn knobs, you're not going to break anything". I encourage each of you to spend time turning knobs and listening to how the sound reacts to each of the settings along the way. There are no studio-police that will fault you for using your equipment how you feel necessary. However, there are some practical limitations to the equipment and how it interacts with audio signals. Always remember that audio is a complex waveform with lots of frequencies at once, some more, or less than others. A tuba can play the same note as a violin, however the frequency difference between these two sound is huge. Due to the complexities of audio, along with our human perspective, most mixes try to keep as much "natural" feeling as possible, while still compressing necessary parts to fit with all the other elements at once. Modern music mixes are hugely layered with stacks, multiple instruments, heavy drums, etc...
Stack-O-Matic
One of the ways engineers creatively use compressors is to apply gradual compression, as opposed to sudden compression. The easiest way to achieve this is to use multiple compressors in your signal chain. Blasphemy! You say. It's more common that you might think. You can wind up with a much smoother, more natural sound by using multiple compressors, each applying light compression, than if you were to try and squash the sound all through a single compressor. In practical application, you will also see engineers combine these compressors with EQ's (equalizers), to help shape the tone of the signal at the same time. The signal flow might look something like this:
The signal chain is: Compressor - EQ - Compressor - EQ. Each instance of compression and EQ are applying slight changes along the way. Start off with a light Ratio setting such as 3:1 or 4:1, and set a Threshold that only captures the loudest peaks in the audio. You'll start to notice that the low frequencies might poke out a bit more due to the compression circuit, that's what the EQ is for. Use the EQ to sculpt out any frequencies that were overemphasized by the compressor. Repeat the same steps with the following compressor, while using the final EQ to take out, or add, any additional frequencies.
While this technique doesn't work for everything, it is a solid approach to elements such as Vocals, which need to stand out in a mix.
Recording
The chain of gradual compression can actually begin during the recording process. Certain elements such as bass guitar, electric guitar, vocals, etc..., tend to sound pleasant when applying small amounts of compression during the recording stage. These elements can be very broad in their dynamic range, from soft to loud within one note. For example, a slight change in the plucking of a guitar can suddenly sound very loud or very soft. The same can apply to vocals, they can be very soft and "breathy" then suddenly change to hard and aggressive. By applying slight compression on the recording chain you can help tame those wild peaks and give yourself a better, more stable foundation. The most ideal way to accomplish this is for your signal chain to go: Microphone - Compressor - EQ - Recording Input. As in the example above, the EQ serves as a sculpting tool to allow you control over the tone before the signal is recorded.
Mixing
Aside from the use of compression on individual tracks, many engineers also choose to use a compressor on the entire mix. This is a personal choice made by each individual. Many studio mixing consoles have a "Mix Bus Compressor" on the main output. You can choose to engage or disengage the circuit based on taste. The main thing to keep in mind when using a compressor on the main outputs is that your entire mix, all the elements, are running through that compressor at once. You don't want to squeeze the mix too hard or it will sound like it's "huffing and puffing" with the drums. Also, keep in mind that you should apply the compressor at the beginning of your mix and leave it engaged while you mix. If you apply the compressor after you've finished your mix there is a good chance it could throw off the balance of the elements as it's compressing them as a whole.
Don't be afraid to experiment and turn knobs. Nobody learned how to use a compressor (or any other equipment) the first time they used it. Especially when working with Pro Tools or other DAW's, you have the ability to make as many changes and try out as many variations as you want. You're only limited by your own imagination. When you feel stuck, try listening to other peoples work and try make your material sound like theirs. The more time you spend with your equipment, the more comfortable you'll be making changes for the big picture.