Putting The Squeeze On: An Intro To Compression, Pt. 2

In part one of this series, we explored the basic principle behind compression, and discussed the various parameters that control how a compressor reacts to an audio signal. In this installment, we will explore using compression in the real world on some of the most common sources you might encounter in your recordings, as well as some general tips and tricks for using compression.


Vocals

The human voice is one of the most dynamic sounds you will encounter in recording, primarily because every possible volume level of the voice is used in modern music, from the softest whisper to the most ear-splitting scream. And even in passages where a singer's volume level should remain approximately the same, it is not uncommon to see level fluctuations of 10 dB or more. Since the vocal is the main point of focus in most types of music, it's imperative that we control these dynamic swings to ensure that the vocal remains at a consistent, audible level, and that certain words or syllables don't disappear behind other elements in the mix.

• The exact compressor settings you use will depend on a number of factors, including the tempo of the song, the individual characteristics of the singer's voice and singing style, and the length of the notes that the singer is hitting. But generally, a good starting place is a fast to medium attack (10ms - 30ms), a fast to medium release (50ms - 200ms), and a 3:1 or 4:1 ratio. Adjust the threshold until you are getting about 6 dB of gain reduction on the loudest peaks.
• If you are trying to keep the voice sounding as smooth and natural as possible, use an optical compressor or any compressor with a soft knee. If you are looking for more "attitude" from the vocal, try a faster FET or VCA based compressor like an 1176, or something with a hard knee. 
• If the vocalist is exceptionally dynamic, you can try using different compressor settings on each song section, using a higher compression ratio, or you can chain multiple compressors together in series. This involves feeding the vocal into a compressor, then feeding the output of that compressor into a second compressor. Try a fast attack and release time with a ratio of 6:1 or higher on the first compressor to shave off only the loudest peaks, and set the second compressor to a slower attack and release and a more moderate ratio to manage the overall level of the vocal. By splitting up the workload between two different compressors, you avoid overcompressing with a single compressor, which can lead to "pumping", "breathing" or other generally undesirable artifacts.
• Tip: Try swapping the position of the compressors in the chain, or using two different models of compressor (a favorite combo of mine is an 1176 into an LA-2A.)

Acoustic Guitar

There are two acoustic guitar techniques that you will probably encounter most often, and each requires a different approach when it comes to compression.

Flatpicking involves using a plectrum (pick) made of plastic, metal, bone, felt or other material to pluck or strum the guitar strings. The percussive action of the pick striking the strings usually causes a fairly pronounced transient at the beginning of each note or chord, but since there is only a single point of contact between the player and the strings, the overall dynamic range of the performance will typically be fairly consistent.

• Try a medium attack with a medium release, and a ratio of 3:1 or 4:1, again shooting for about 6 dB of reduction on the loudest peaks. 
• For strummed rhythm parts, identify the primary note value being used (quarter note, eighth note, etc.), then set the release so that the compressor engages and returns to 0 dB of reduction within one instance of that note value. This can create a sense of motion that will match perfectly with the groove of the song and help hide the compression effect.
• If the attack of the guitar is too present, a faster attack time on the compressor can smooth the sound and help it to blend into the mix. You will likely need to reduce the threshold, however, to avoid overcompressing.

Fingerstyle playing involves plucking the guitar strings with the finger tips or the fingernails. This generally results in softer transients, but because the player is using several fingers simultaneously to sound the different notes, volume levels can vary dramatically from finger to finger or chord to chord, creating potential problems when trying to compress the sound. The context of the part should guide your decision-making process when choosing compression settings. If it is a solo part, be very careful not to overcompress, as the natural dynamics of the part will add a sense of realism and depth to the recording. If the part is surrounded by other instruments, however, it may be necessary to use more compression to ensure that the individual notes can be heard in the mix.

• One of my favorite ways to use compression on arpeggiated fingerstyle parts is to subtly increase the sustain between the notes, making the part sound more full and "up-front". This requires very careful attention when setting the attack and release parameters though, so as not to squash the delicate transients or introduce pumping artifacts. Try a medium attack (30 ms or so), a fast release (try the fastest release time your compressor will allow and increase it as needed), and a ratio of 2:1 or 3:1, and aim for gain reduction of around 2-3 dB.
• Tip: For this and other applications requiring a fast release, try to avoid using optical compressors. They are renowned for their smoothness due to a softened attack and release inherent to the design, but this makes it very difficult to achieve fast release times without introducing artifacts. Try using a VCA, FET or variable-mu design instead.
• Fingerstyle playing has a tendency to introduce more low-end energy than flatpicking, and the added energy can cause the compressor to respond with more gain reduction than you would like. To get around this, try sending a copy of the guitar to an aux track and using a high-pass filter to remove everything below 200 Hz or so, then send the output of the aux track to the sidechain input of your compressor. This is essentially telling the compressor to ignore the lower frequencies and compress based only on the level of the mid and high frequencies.
• A sidechain is a special input found on some compressors that will allow the compression circuit to be controlled by a signal other than the one being fed into the compressor's main input. The two main uses of the sidechain are frequency dependent compression and ducking.

Electric Guitar

Aside from synthesizers, there is probably no instrument in modern music that is more versatile than the electric guitar. It is capable of producing an enormous range of wildly differing tones and textures, and this chameleon-esque quality allows it to fit into just about any musical context or style imaginable. And the electric guitar is no stranger to compression; in fact, compression is the foundation of many common guitar sounds. Amplifier distortion, for example, is created by pushing gain stages within the amp to the point of hard clipping. This clipping naturally compresses the dynamic peaks of the signal, and is what gives distorted guitars such a huge, in-your-face sound. But compression can also be used much more subtly on clean or even slightly distorted electric guitar to add sustain and to smooth out a performance.

• For clean and punchy electric guitar, try a medium attack speed and a fast to medium release with a ratio of 3:1 and adjust the threshold to taste.
• For distorted guitar, try a fast attack and a medium release, aiming for about 4 dB of gain reduction.
• To get a smoother and more sustained clean or slightly dirty sound similar to the one heard on the intro of the Red Hot Chili Peppers' "Under The Bridge", try a fast attack time with a medium to slow release.

Piano

The piano is one of the more challenging instruments to record and mix. Its naturally dynamic nature and wide frequency range can cause it to easily dominate a mix one moment, and all but disappear the next. When it comes to compressing the piano, your decisions will depend greatly on the musical style and the context in which the piano appears.

• If the piano is by itself or within a sparse arrangement, try using less compression so as to not destroy the natural dynamics of the performance. This is especially important in jazz and classical styles, where the player may utilize the entire dynamic range of the instrument within the same piece to create changing moods and textures.
• If the piano is being used in a dense arrangement, try using a medium attack, a medium to long release, and a 4:1 or 5:1 ratio. This will allow the initial attack of the hammers through, but will attenuate the sustain of the chords to keep them from muddying up your mix. Don't be afraid to really push the compressor hard to get 10 dB or more of gain reduction.

Bass Guitar

In most styles of music, the bass guitar (or synth bass, stand-up bass, etc.) serves as the harmonic foundation of the entire arrangement. Because of the amount of energy being created by the low fundamental frequencies, the listener can often feel the bass line as much as they can hear it, so any fluctuations in volume will be especially noticeable and the bass dropping out unexpectedly can rob the music of power and focus. For this reason, it is very common to  compress the bass much more than you would other instruments, with the idea being to "cement" it in place so that it occupies the same sonic real estate throughout the song.

• For bass lines that contain mostly half notes or whole notes, try a fast attack with a long release (600ms - 700ms) and a ratio of 6:1. Don't be afraid to apply 10 dB or more of gain reduction. The goal is to remove as much of the dynamic range as possible.
• For busier bass lines, such as walking lines or slap-pop style, try using a significantly shorter release (200ms - 300ms) to avoid the compressor "hanging over" between notes. Also, try a longer attack time to allow the percussive nature of the attack to come through and add rhythmic energy to the track.
• Be very careful when setting the attack on the compressor, as very fast attack times can cause unpleasant distortion when compressing low frequencies. The reason for this is that it takes longer for low frequencies to complete one full wave cycle (around 10 ms at 100 Hz), so if the attack is set too fast, the compressor is actually altering the shape of the bass's wave form, causing it to distort.
• Be very careful to avoid pumping or breathing artifacts, especially if they are out of time with the track. The extra energy in the bass frequencies will make this very noticeable, and it can quickly wreck the balance of your entire mix.
• If you find that the bass guitar is masking the kick drum and you are unable to fix the problem using only EQ, try adding a second compressor to the bass track and triggering it from the kick drum using the sidechain input, aiming for 2 or 3 dB of gain reduction. This will effectively "turn down" the bass momentarily each time the kick drum hits. This is called ducking, and it's the same technique that commercial radio stations use to turn the music down automatically anytime the DJ is speaking.

Drums

The way that compression affects drum tracks is dependent on a number of factors. Compressors will react differently to a snare drum than to a kick drum, and differently to a room or overhead mic than to a direct mic. The type of compressor you are using also plays a huge role. Optical compressors are generally too slow to keep up with the lighting fast transients created by percussion instruments, and can have a tendency to "smear" the sound of the drum or cause unwanted pumping artifacts. Generally, it's best to choose a compressor with a FET or VCA design, as these tend to have a much faster response. Also, it's very important to pay close attention to the attack setting; too fast of an attack time will squash the initial transient and rob the drum of punch and power, while too long of an attack time, and the compressor won't act quickly enough to grab the decay of the drum.

• For "standard" compression on a direct-miked drum, try starting with a ratio of 4:1 and an attack time of 15ms or greater, with a fast release. Adjust the attack and threshold until you are getting the amount of control that you want.
• A great way to enhance the sound and relative level of the room mics is to process them through a compressor set to a high ratio and a very fast release time with lots of gain reduction (usually 10 db or more). This technique can be be used to a wide range of effects, from just adding a little more attitude, to increasing the "bigness" of the snare drum, to total sonic destruction. (This works especially well on a Universal Audio 1176 or an Empirical Labs Distressor.)
• If you find that you are hearing too much snare drum in the overhead mics, try inserting a compressor on the overheads and side-chaining it from the direct snare mic. Adjust it until the presence of the snare is reduced in the overheads without audible pumping in the overheads (usually 3 to 4 db of gain reduction). A fast release will help hide the effect.