CHAPTER 8
Shaping Dynamics
The various mastering compressors have such different qualities. Some are loved for their punch, warmth, vintage tone, smoothing, general character, or maybe the way they treat high frequencies. Their strength also might be their functionality, to tame dynamics, and to minimize the effects of limiting. Every compressor is different and usually takes time to “get.” Of course, it would take much more than a chapter to fully describe the nuances of every mastering compressor, so the common ideas are described here.
Compression is actually far less of a factor in today’s professional mastering than most beginners would believe. There is so much hype surrounding compressors that it misleads many people about the significance of their role.
Of course, compression as we normally think of it is not the only dynamics processing used in mastering. There are several, and each type is discussed in this chapter.
Types of Dynamics Processing
There are several types of dynamics processing used in mastering, including:
• Compression (also called downward compression)
• Expansion (also called upward expansion)
• Parallel compression (also called New York compression or upward compression)
• Side-chain compression (alternate signal triggers compression action)
• Multiband compression (compression of specific frequency ranges)
• Limiting (a type of compression with a very high ratio)
Compression/Downward Compression
Downward compression is the most common dynamic range processing in mastering. Some mastering engineers rarely, if ever, use any other type, except limiting. With downward compression, the loudness of a recording is lowered when it exceeds a set threshold. Separate from compressing the dynamic range, there are other side effects called character.
Expansion/Upward Expansion
Expansion is when the loudness of a recording is raised when it exceeds a set threshold. It is rarely used in mastering, although some engineers may choose to use it with a recording that is overly compressed during mixing.
Parallel Compression
Parallel compression is performed by mixing a compressed signal in with its original. Parallel compression is used to effectively raise the lowest sounds. This process is usually best if applied subtly. It has been gaining in popularity in mastering and mixing. Some professional mastering engineers use it, whereas many prefer to use only downward compression.
Parallel compression can be accomplished manually by making a copy of a recording, ensuring that the original and copy are absolutely perfectly aligned. The copy then can be compressed and mixed in with the original. This should be done very subtly and can easily be as low as –20 dBFS or more below the original track, depending on the recording at hand. Parallel compression can be accomplished automatically by using the “Mix” knob on many plug-ins and some hardware units. The “Mix” knob is usually measured in percentages, with 5 to 20 percent often being the full range of possibilities for this technique in mastering. Because parallel compression is mixed in so subtly, it can involve extremely low thresholds. Of course, it should be set according to what pleases the ear.
Parallel compression is especially popular with acoustic or orchestral music and when reverb or low-level signals need to be raised.
Side-Chain Compression
Side-chain compression is when a signal is used for triggering the compressor action other than the signal being compressed. In mixing, side-chain compression is often used to compress one track according to the dynamics of another track.
In mastering, side-chaining is used a different way. The compressor will be triggered by an equalized version of the track being compressed, usually with bass frequencies de-emphasized. This is to avoid the compressor overly responding to the bass frequencies, which have the most energy and otherwise would dictate the compression action. Side-chaining is a somewhat popular option in mastering, although some professional engineers never use it or feel that they need it.
Multiband Compression
Multiband compression divides a signal into several frequency bands and applies compression to them separately.
Limiting
Limiting is essentially compression with an extremely high ratio—it is used to set a ceiling (a maximum level) that nothing should exceed. For the most in dynamics and depth, no limiting is used. When achieving loudness is a goal, limiting is virtually always in use.
Limiters impart their own character on the sound. Because of this, it is important to discover a limiter that works well with the other devices in the chain. Today, digital limiting, with look-ahead accuracy and high quality, is the standard. Digital limiters such as the Fabfilter Pro-L, Voxengo Elephant, and Slate Digital FG-X are popular. In the modern mastering studio, analog limiting has all but disappeared.
Compression Settings and Meters
Virtually every mastering compressor has four primary settings. These settings include the following:
• Threshold. A threshold is the level above which a compressor begins its action.
• Ratio. A ratio defines how much action will occur when the level goes over the threshold.
• Attack. The attack is how long the compressor takes to achieve its action.
• Release. A release is how long the compressor takes to end its action once the level has gone below the threshold.
Setting Attack/Release/Threshold
The settings of a compressor depend on the compressor itself and the recording at hand. As mentioned earlier, it takes time to “get” a compressor.
One general approach is to find settings that produce the most pumping, then set the attack and release times to work with the music, and finally, roll the threshold and ratio down to set the effect in with subtlety. A release time of 250 to 300 ms is often a great place to start. However, many engineers work with compressors much more intuitively than this approach.
Transients are an abrupt or sudden change in level, for example, we might think of a drum sound such as a snare as being very transient. Compressors affect transients when they are set to a fast-enough attack time. Slow attack times minimize the effect on transients. Negative effects on transients are usually easiest to hear when listening to the drums. If a compressor is having a negative impact on drum sounds, the attack time can be raised.
In mastering, a ratio of 1.5:1 is quite typical, while a ratio of 4:1 would often be considered high. An attack time of 0–10 ms is often considered fast, while an attack time of 30 ms might considered medium or slow. A release time of 100 ms would be considered very fast, while a release of 250–300 ms is quite typical. Compression settings depend on the material and task at hand, so more exacting compression recipes are avoided in this book.
Compressors are best set with a goal in mind, and achieving the envisioned goal may or may not involve any specific technique.
Macrodynamics/Microdynamics
The term macrodynamics refers to loudness over a long passage, primarily entire parts of a song, such as an intro, chorus, or verse. Microdynamics are faster sounds than macrodynamics. You might think of a drum sound or other fast sound. Microdynamics are mainly the transients and peaks, which have a much shorter duration. Both macrodynamics and microdynamics are considered in mastering.
RMS-Sensing Compressors
Root-mean-square (RMS) and peak sensing are two different methods compressors use to detect levels. Compressors that react to the average level are called RMS-sensing compressors and also may be called averaging compressors. The RMS/average level is more akin to how we actually perceive loudness when listening than peak level. The sound of RMS-sensing compression is usually more controlled. Because of this action, these compressors work best for macrodynamic compression.
Peak-Sensing Compressors
Peak-sensing compressors react to peak levels. Limiters, which are essentially compressors set with a very high ratio, are usually peak-sensing because they must quickly track the highest levels of a recording. Peak-sensing compressors are much more noticeable and less graceful than RMS-sensing compressors.
Compressor Response
Two different brands of compressors will produce different results, no matter whether they are both peak- or RMS-sensing—even with the same settings.
Character versus Transparent Compression
A processor that has character is one that imparts some tonality, as opposed to a transparent processor, which performs its function with little or no influence on the sound. Character compressors sometimes add their tonality even if no dynamics processing takes place. Other times character is imparted by various levels of processing. One must learn a compressor’s nuances to skillfully apply its character. This is something that comes in time after experience with a compressor. Also, it is often discussed in forums, may be discussed in the user manual, or can be explored with a cordial call or e-mail to the manufacturer.
Transparent compression provides dynamics processing with very little harmonic distortion. Some units are capable of both character and transparent compression.
Mixbus Compressors
Mixbus compressors are used by mixing engineers on their master bus. Often they mix with the compressor on the bus during the entire mixing session. Mixbus compressors are usually ones that impart a considerable amount of character, whereas mastering compressors are generally more transparent. Sometimes compressors that are typically thought of as mixbus compressors are used in mastering for recordings that need more character.
Punchy Compression
This type of compressor can have a huge impact on the sound, and there are many factors that can cause this impact. For instance, the type of transformer can have a tremendous influence on tonality. Circuit architecture can make it punchier or less punchy. For example, an API 2500 is great for fast or aggressive music and adding punch while also remaining quite versatile. In contrast, the Pendulum ES8 provides a very hi-fi sound but is less punchy and is not as fast.
Serial Compression
Serial compression is the use of two compressors. Sometimes the purpose will be to combine the characters of several units. Also, different compressors have different strengths, and sometimes combining two can be just the right fit. For example, one compressor might be set for microdynamics processing, whereas another is set for macrodynamic processing. It is very common to use one compressor to tame peaks, with a fast RMS- or a peak-sensing action, along with another that is used with slower settings and with an RMS/averaging sensing action.
Volume Automation for Macrodynamic and Microdynamic Adjustments
Sometimes, instead of using a compressor, volume automation can be used. In many situations, this can provide a more transparent adjustment than using a compressor. Whether working to fix a specific microdynamic or macrodynamic problem, this can be the best option. Remember, the action of a compressor is intended to be an automated loudness adjustment—it also can be done manually. For example, if a chorus is too loud in comparison with a voice, instead of working with an RMS-sensing compressor with slow attack and release times to even out the difference, a simple volume-automation curve could be used. Also, if there was a kick drum within a song that accidentally rose high above the rest of the recording, if a remix is not an option, a simple volume-automation curve might produce better results than a compressor.
Gain-Reduction Meters
Most compressors have a meter that shows the gain reduction taking place. In mastering, often compression is used subtly. Gain-reduction meters may not show any or only very slight movement, even when the compressor is optimally set.
Compressor Input Levels
Some compressors, such as the Manley Vari-Mu, have an input control that can have a big impact on the sound. With the Vari-Mu, if the input level is raised compared with the output, the tubes are driven harder, and more character is added.
Compressor Output/Makeup Gain
The output/makeup gain can be used on a compressor so that the input and output signals are at the same level. In this way, the effect of the compressor can be auditioned using the bypass feature. It also avoids the difference in loudness affecting the engineer’s perception of the processing. Makeup gain also can be used to simply add gain into the signal. This must be done carefully because the makeup gain of different compressors can sound much different.
Linked versus Unlinked Compression
Linked and unlinked modes on a stereo compressor indicate whether the compressor will have the same or different action on each channel. Linked compression mode allows the compression to be more apparent, which may be useful when a compressor has a certain character. You should always try linked and unlinked modes to discover which sounds better.
Expansion/Expansion Before Compression
A few engineers will use expansion processing to expand the dynamic range or use a transient designer process to highlight the transients and increase the dynamic range of a recording before using a transparent limiter to compress it again. This technique is not very popular or well accepted, although it is intended to add punch and expand the sound stage.
Analog Compressors
There are several different types of analog circuit methods used for compression. The types include voltage-controlled amplifier (VCA), opto/ELOP, variable mu, field-effect transistor (FET), and pulse-width modulation (PWM). Each of these gain-control mechanisms has its own characteristics.
Voltage-Controlled Amplifier (VCA)
The term VCA compressor is used when integrated circuits (ICs) are used to control the gain reduction. Technically speaking, virtually all analog compressors could be called VCA compressors because voltage is being controlled, but in studio parlance, it refers only to ones that use ICs. VCA compressors are known for being very fast and are used on the fastest-tempo rock and metal, along with many other styles, and can work with slow tempos as well.
Opto/ELOP
Opto compressors used for tracking and mixing are quite slow. Opto compressors created for mastering are much faster, including the Pendulum OCL-2 and the Shadow Hills Mastering Compressors’ opto side. Opto compression is normally talked about as having a smoothing effect, especially smoothing the high frequencies, with lots of “mix glue” while adding a subtle amount of fullness. Their action is based on a light-controlled optical element to produce smooth and pleasing gain reduction.
Variable Mu
Variable mu compressors are thought of as having more character, as well as providing smoothness, depth, and fullness. With this type of compressor, gain reduction is accomplished with a vacuum tube.
Field-Effect Transistor (FET)
FET compressors are not as widely used in mastering. Often FET compressors are thought to impart too much tonality for mastering.
Pulse-Width Modulation (PWM)
PWM compression is typically used to emulate the types of compression listed earlier.