Chapter 4

Fundamental Mastering Tools and The Primary Colors of Mastering

I’ve reviewed how an effective mastering studio represents a symbiotic component of the entire mastering system. Next, we will focus on the equipment used to make the tonal or surgical adjustments and sonic enhancements. These include the following five fundamental tools for audio adjustments: analog-to-digital and digital-to-analog converters (AD/DAs), equalizers (EQs), compressors/limiters, expanders, brickwall limiters (BWLs), and digital audio workstations (DAWs). I refer to three items on this list—EQs, compressors and BWLs—as The Primary Colors of Mastering due to their seminal importance in audio mastering. In this chapter, I will review the function and desirable attributes of these fundamental mastering tools.

Analog-to-Digital and Digital-to-Analog Converters (AD and DA)

As their names suggest, these devices convert audio stored in the analog (signal-based) domain to the digital (sample-based) domain, and vice versa.1 Digital audio must first be converted to analog signal to be heard or used with analog equipment. At minimum, a professional mastering studio requires two or three DA converters and one AD converter. Two DA converters are needed for the PBDAW—one for monitoring flat or unprocessed source audio, and another for the channel path to make adjustments and enhancements to the audio. A third DA is needed to monitor the final processed audio from the RDAW. If you use a modern mastering console that features a mult (duplicate) of the input signal, then you only need two DA converters for the mastering system, as the PBDAW signal can be accessed after the mult for both the flat audio monitor position and the channel path. Additionally, the system requires one AD converter to convert the signal from the analog processing chain to digital and then into the RDAW.

Ideally, a mastering-quality DA converter will present audio that is focused and accurate without adding coloration, distortion or enhancements. It should be able to convert all common digital audio formats2 and resolutions to analog transparently. A mastering-quality AD converter will also accurately sample and digitize the analog signal into the digital binary code—free from clock jitter and aliasing filter artifacts—at all common sampling frequencies and bit depths. It should preserve transients and dynamic range exactly as they are presented by the analog signal.

Mastering Reference Levels

DA and AD converters (Figure 4.1 Figure 4.2 Figure 4.3 Figure 4.4 Figure 4.5) usually include onboard level adjustments for calibrating the chosen operating reference level for the mastering studio. This calibration correlates digital decibels full scale (dBFS) levels with analog voltage readings and the VU meter so that lower dBFS levels (adjusted at the DA converter to 1.23V = 0VU = +4dBu) raise quieter more dynamic source files for a more robust level through the mastering system. The reference level calibration is accomplished by playing a 1kHz tone at the PBDAW at the selected dBFS level (usually −14dBFS), verification of voltage level at DA conversion (1.23V), then to the AD converter, the RDAW and post RDAW DA converter for end-of-chain monitoring. Reference level calibration is critically important for establishing headroom (region of signal between a nominal level and clipping—often 0dBFS digitally and +24dBu analog) in a mastering system, and is usually calibrated as one of three options (Table 4.1).

Figure 4.1

Figure 4.1The Lavry Gold AD122–96MX is an exemplary mastering AD converter. It is sonically transparent and features low jitter, musical clipping properties near 0dBFS, soft limit and soft saturation functions for loudness enhancement, full scale metering, and onboard test tones.

Source: (courtesy Lavry)

Figure 4.2

Figure 4.2The Lavry Quintessence DA converter with onboard monitor control.

Source: (courtesy Lavry)

Figure 4.3

Figure 4.3The Lavry Blue 4496 chassis with one AD converter and three DA converters. These converters are also feature-rich and work great in a mastering context.

Source: (author collection)

Figure 4.4

Figure 4.4The Prism ADA-8XR mastering-grade AD and DA converter with monitoring/headphone functionality. A modular system available in multiple configurations, this unit has the 8AD plus 8DA setup. Also has the ‘over-killer’ limiting option for creating louder masters.

Source: (courtesy Capitol Mastering)

Figure 4.5

Figure 4.5The Crane Song HEDD (Harmonically Enhanced Digital Device) features high-quality AD conversion with DSP-modeled analog/tube coloration (triode tube, pentode tube, and analog tape compression options).

Source: (courtesy Crane Song)

Table 4.1Shows the relationship between a 1kHz tone dBFS level, its output in voltage at the DA converter, the aligned voltage at the DA converter, and the VU meter alignment. Represents an overview of reference level calibration of the mastering studio. The Dangerous Convert-2 has a preset level calibration selector for these three common options (see Figure 3.7).
Mastering Studio Reference Level Calibration
dBFS
(1kHz at RDAW)
Volts
(at DA/Adjusted DA)
VU Cal dBu
−18 .775/1.23 0 +4
−16 .975/1.23 0 +4
−14 1.23/1.23 0 +4

I have always used −14dBFS = 1.23V = 0VU = +4dBu as my mastering reference level. Most mastering studios are similarly referenced, especially when working in modern pop, rock or hip-hop genres.

Equalizers (EQ)

An EQ is one of the three critical primary colors of mastering. EQ allows you to boost or cut selected frequencies and bandwidths. As musical instruments possess their own range of frequencies, this tool allows the Mastering Engineer to either feature—or draw attention away from—an instrument or group of instruments. Larger bandwidth or shelving curves are utilized for tonal balancing across larger areas as needed, and narrower bandwidths are useful for detailed or surgical moves. The power of equalization to enhance the fidelity and presentation of recordings remains a profound component of audio mastering.

Equalizer Configurations

There are various types of EQ—filters, shelving, graphic, fixed parametric and parametric are most of the EQ types you will encounter as hardware devices or plug-ins.

Filtering EQs either remove frequencies above a selected frequency—known as a low-pass filter (LPF), or below a selected frequency—known as a high-pass filter (HPF). The slope of the filter can be adjusted in a dB/octave ratio, with 6dB/octave representing the most gradual slope and 24dB/octave the sharpest. The filtering slopes are in multiples of 6dB, with each of four options referred to as an order: first order is 6dB/octave, second order is 12dB/octave, third order is 18dB/octave, and fourth order is 24dB/octave. These are excellent for managing high-frequency issues or low-frequency rumbles (necessary for vinyl cutting), and are also useful in conjunction with shelving EQs to keep boosts from becoming excessive and causing a tonal imbalance in the high or low frequencies.

A shelving EQ boosts or cuts all frequencies above or below a selected or fixed frequency, making it suitable for tonal balancing of highs or lows. Baxandall EQs are a shelving EQ curve adopted and mass-marketed as the treble and bass controls on consumer stereo receivers. They have seen a resurgence of interest from Mastering Engineers and professional audio manufacturers alike (i.e. the Dangerous Bax EQ) for a natural and familiar tonality. A resonant shelf is a unique shelving curve attributed to Michael Gerzon3 that dips at the shelving frequency and never plateaus—making it excellent for adding high or low-frequency extension, especially in M/S to bring vitality to a lifeless vocal (see Chapter 15—Mid-Side). This type of EQ can be simulated with a wide band on a parametric set for a frequency below 20Hz or above 20kHz, resulting in the same effect, otherwise known as the left or right half of a wide bell curve. A tilt EQ will shift along a selected frequency ‘fulcrum’ and gradually boost or cut to either side of it in a very gradual manner—best applied for tonal shaping.

It is worth mentioning one of the holy grail EQs, the Pultec EQP-1A. The Pultec is a passive EQ (using passive electronic components—resistors, capacitors, and inductors—that are not powered until output stage amplification) designed and produced beginning in 1951 by Ollie Summerland and Gene Shank. They can be set for rich low-frequency enhancement and extended shimmer in the high frequencies. There are only two selectable frequency bands and a low-pass filter, but the ability to boost and attenuate simultaneously creates unique and coveted EQ curves that many manufacturers have since endeavored to duplicate. Pairing the EQP-1A with the MEQ5—a three mid-band EQ—allows for more flexibility. Figure 4.6 shows the vintage pair of EQP-1As from Capitol’s Studio B.

Graphic EQs have a set series of frequencies at a fixed bandwidth that can be cut or boosted via sliders. These are not common in a mastering context, and are often used for room tuning or live sound (PA) reinforcement applications.

Fixed parametric EQs (example: NTI EQ3 [Figure 4.7] or Maag EQ4M with only HF ‘air’ shelf selectable) do not allow for the adjustment of bandwidths or frequencies—likely contributing to less expensive manufacture and quick implementation for the user. They have practical applications for mastering, but frequency limitations require using additional EQ options in a mastering system.

Figure 4.6

Figure 4.6A pair of Pultec EQP-1A EQs from Pulse Techniques in Englewood, NJ. Pultecs are passive EQs—they don’t require power in the EQ circuit and use passive components (resistors, capacitors, and inductors) followed by a tube gain makeup amp.

Source: (courtesy Capitol Studios)

Figure 4.7

Figure 4.7An NTI fixed parametric EQ. Stepped attenuators at 3dB clockwise steps, and ¼dB counter-clockwise steps for precise adjustments. ‘Air band’ is popular in stereo buss/mastering applications.

Source: (courtesy Capitol Studios)

For mastering applications, the parametric equalizer remains the primary EQ of choice due to its flexibility in selecting bandwidths (Q)4 and frequencies. The design and concept is credited to Burgess Macneal and George Massenburg,5 who designed EQs under the ITI,6 Sontec, and then later, GML names. This tool allows for precision selection of the independent parameters (bandwidth, boost/cut, and frequency) that will accentuate or attenuate the chosen frequency (or frequencies). To further inform your mastering acumen, it is helpful to familiarize yourself with the following seminal parametric EQs: ITI ME-230 (Figure 4.8), Sontec MEP-250C and MES-432C (Figure 4.9); GML 8200 (Figure 4.10) and 9500; Pultec-style or Enhanced Pultec EQs—Pulse Techniques’ EQM-1A3, EQM-1S3, and MEQ-5, Manley Labs’ Enhanced Pultec and Mid EQ (Figure 4.11), and the Bettermaker Mastering EQ; Maselec MEA-2; Tube-Tech EQ 1AM; and Weiss EQ1 (Figure 4.12), an acclaimed digital hardware parametric. Notice these mastering-grade parametric EQs generally include the common functionality of a high-pass filter (HPF), low-pass filter (LPF), high shelf (HS), low shelf (LS) and three or four parametric EQ bands for both tonal shaping and surgical EQ approaches. Additionally, they are fitted with stepped attenuators with precision resistors7 for accurate recall and reliable left/right stereo imaging. Any coloration differences between them are due to discrete components and circuitry design, bandwidth options, the number of EQ bands, frequency options, cut/boost increments, and range.

Figure 4.8

Figure 4.8The first commercially available parametric EQ—the ITI ME-230 circa 1971.

Source: (courtesy Capitol Studios)

Figure 4.9

Figure 4.9Burgess MacNeal started Sontec in 1975 from remnants of defunct ITI, and produced one of the quintessential mastering EQs. Pictured here is one of two Sontec MES-432Cs I use daily at Capitol.

Source: (courtesy Capitol Mastering)

Figure 4.10

Figure 4.10George Massenburg’s first post-ITI product, the GML 8200 (the GML 9500 is the mastering version with stepped attenuators in 0.5dB steps and precision resistors).

Source: (courtesy Capitol Studios)

Figure 4.11

Figure 4.11Manley’s impressive take on producing mastering-specific Pultecs with stepped attenuators and precision resistors circa 1990–1993. These EQs impart a rich tonality, smooth high-frequency extension and offer the sought-after Pultec EQ curves.

Source: (author collection)

Figure 4.12

Figure 4.12The Weiss EQ1 is a seven-band digital hardware parametric EQ. It comes in four configurations (basic, linear phase, dynamic and dynamic-linear phase) and includes M/S functionality.

Source: (courtesy Weiss)

With the advent of plug-ins, the adjustable settings of a parametric EQ become virtually unlimited, permitting detailed adjustments not feasible in their analog counterparts. These user parameters may include options such as M/S for each band, parallel or series interaction between bands, and phase linear, minimal phase, and analog phase options. Excellent plug-in parametric EQs worth experimenting with are the DMG Equilibrium (Figure 4.13) and the FabFilter Pro-Q3 (Figure 4.14). This precision adjustment of parameters is true for all plug-in audio processing equipment including compressors and limiters—hence my assertion that the best modern mastering system incorporates both digital and analog domains (see Chapter 14—Advanced Mastering Tools).

Compressors/Limiters

These devices manage dynamic range and transient peaks by lowering (compressing) program level above a selected input threshold. They can also be used to add apparent volume or loudness by then adding output gain to the compressed audio. They are used in mastering to keep audio signal from distorting, add sonically pleasing coloration, increase the RMS level of program material (loudness), and create a hyped ‘radio’ sound. A standard downward compressor will compress or lower peaks above a selected input threshold. However, if you use a compressor in parallel with the main program signal, the combination functions as an upward compressor, bringing the ‘valleys’ up. This is handy for adding detail, clarity, and apparent volume while still preserving transient response, thus maintaining vitality and punch (see Chapter 14—Advanced Mastering Chain Tools and Techniques). Perhaps the purest form of upward compression is to raise low sections of a recording with level automation or editing, leaving the transient peaks alone.

Figure 4.13

Figure 4.13The DMG Equilibrium EQ plug-in is feature-rich with multiple filter, shape, phase control, configuration, analysis, and interface display options. It models a hall-of-fame selection of coveted vintage analog EQs (that can be operated linear phase!) including Pultec, SSL E and G series, Sony Oxford, Focusrite ISA 110, API 550, Neve 88, Harrison 32 C, Sontec 250, and GML 8200. Each band can be used in M/S mode. I use it regularly—the ‘desert island’ plug-in EQ.

Source: (courtesy DMG)

Figure 4.14

Figure 4.14FabFilter Pro-Q also boasts a great user interface and impressive features such as phase linear operation, dynamic EQ, M/S processing, and spectrum analysis. The EQuilibrium doesn’t have dynamic EQ, but the Pro-Q doesn’t have vintage EQ modeling.

Source: (courtesy FabFilter)

Compressors typically have adjustable controls for input, threshold, ratio, attack time, release time, and output. A brief description of each follows: input controls the input level, threshold determines the level after which compression will begin, ratio determines the amount of compression (i.e. a ratio of 4:1 means that for every 4dB of input level above the threshold, the unit will output merely 1dB), attack and release times refer to the speed of compression action and its subsequent release above the threshold, and output refers to the output gain (sometimes indicated as makeup gain) added after compression.

By contrast, a limiter stops or limits transient peaks in program material by using a much higher ratio than does a compressor. Also, the attack and release settings are set to more intensely respond to program above the threshold setting. For analog compressors with limiting functionality, such as the Manley Variable-Mu™ or Smart C2, this limiting involves selecting from the higher ratio settings and adjusting the threshold for action on loud peaks, usually with medium to slow attack times and fast release times. In figurative parlance, a limiter is the ‘bigger and hairier’ version of a compressor. A limiter can prevent downstream equipment from distorting; or if the engineer raises the overall gain feeding the limiter, the average root mean square (RMS) level increases, resulting in louder sounding masters. For example, the Smart C2 has an ‘L’ ratio setting for limiting, the SSL Stereo Compressor has a ratio setting of 10:1, and the Urei/Universal Audio 1176 has 20:1 as a limiting ratio setting. A related note on the 1176—most engineers know of the ‘all-buttons’ mode that is revered for its renowned ‘pumpy’ and overdriven sound. I will continue this chapter by primarily discussing hardware compressor/limiters, but most of these units have meticulously modeled plug-in counterparts for digital applications.

Varieties of Analog Compressors

Analog compressors are usually designed utilizing one of four different gain reduction-based circuits: electro-optical, field effect transistor (FET), variable-gain, and voltage-controlled amplifier (VCA). A fifth type, a diode-bridge compressor, is less prevalent, but is used in the Neve 2254 and 33609 compressors (see Figures 4.32 and 4.33). Bear in mind that the sound of compressors is also informed by elements beyond the gain-reduction circuit. Transformers or tubes in the circuit design and user-controlled features result in vast sonic differences, so consider all aspects when selecting or auditioning a mastering compressor.

Electro-optical compressors are smooth and musical, and produce warm coloration. They perform gain reduction by virtue of a photocell that reads the brightness of a bulb, light-emitting diode (LED) or electroluminescent panel which is correlated to input level. The classic electro-optical compressor/limiters from the 1960s and 1970s—the Universal Audio Teletronix LA-2A (Figure 4.16),8 LA-3A (Figure 4.17) and LA-4A—only have an input and output adjustment so that threshold, ratio, attack, and release are internally set. The time constants for attack and release remained non-linear meaning program dependent response (fast then slow release, for example) and create a slow or ‘spongy’ response characteristic. This makes these compressors well-suited for tonal enhancements and slight peak management. As the demand for stereo buss compression increased, manufacturers added features and functionality to the coveted classic designs to offer stereo electro-optical compressors made specifically for buss compression or mastering such as: the Manley SLAM!, PrismSound Maselec MLA-2, Tube-Tech CL2A, Pendulum OCL-2, and Shadow Hills Mastering Compressor (Figure 4.19).

Figure 4.15

Figure 4.15Gain reduction circuit component types, from left: the Teletronix LA-2A’s renowned T4A optical attenuator with electroluminescent panel and photo resistors, a vactrol light dependent resistor (LDR), voltage-controlled amplifier (VCA), field effect transistor (FET), and vacuum tube.

Source: (courtesy Ian Sefchick)

Figure 4.16

Figure 4.16Universal Audio Teletronix LA-2A electro-optical compressor. Originally designed by Jim Lawrence in the early 1950s. Revered for its musical multi-stage release time characteristics, it inspires many modern electro-optical mastering compressor designs.

Source: (courtesy Capitol Studios)

Figure 4.17

Figure 4.17Universal Audio/Urie LA-3A electro-optical compressor pair. Uses the same T4 optical attenuator as the LA-2A, but with solid-state electronics (tubes were considered old technology by the late 1960s).

Source: (courtesy Capitol Studios)

FET compressors utilize a transistor to perform the gain reduction on the program material. They offer greater versatility and control on transient peaks than an electro-optical compressor since threshold, ratio, attack, and release are user-adjustable. Characteristic is a focused or intimate type of coloration increasing with extreme gain reduction settings. Classic examples are the Urei 1176 (Figure 4.20), 1178, and 2–1176.9 A FET compressor can be quickly set up to function as a limiter by increasing the ratio to the highest setting and then carefully adjusting the threshold (generally higher), attack and release settings for the desired peak reduction response. Modern units suitable for mastering are the Manley SLAM! (Figure 4.21), Crane Song STC-8 (Figure 4.22), and Overstayer 3706 SFE.

Figure 4.18

Figure 4.18A pair of Summit Audio TLA-100A compressors. These are often misconstrued as electro-optical compressors but actually use a proprietary VCA gain reduction circuit.

Source: (courtesy Capitol Studios)

Figure 4.19

Figure 4.19The Shadow Hills Mastering Compressor. This unit boasts two compressors in series: an electro-optical compressor first, followed by a discrete VCA compressor. The output transformer is selectable between nickel, iron, and steel for different coloration options.

Source: (courtesy Shadow Hills)

Variable-gain compressors utilize the re-biasing of a vacuum tube to handle the gain reduction duties. This topology does not have a user ratio control, as the ratio increases along with the input amplitude. These compressors are generally known for a warm or smooth coloration characteristic and provide tonality more than the quick control of peaks. Of course, each unit must be carefully listened to for coloration assessment—one of my variable-gain compressors at Capitol imparts a pleasing high-frequency extension along with the expected ‘glued-together’ component, whereas other versions of the same compressor do not. Attack times are quicker than an electro-optical compressor, but still not at the peak-managing functionality of a FET or VCA compressor.

Figure 4.20

Figure 4.20Universal Audio 1176 compressor pair. Released in 1968, the 1176 implements a FET gain reduction circuit and is known for fast attack and release time settings, and range of tonal characteristics from clean compression at the 4:1 ratio to compelling saturation in ‘all-buttons-in’ mode.

Source: (courtesy Capitol Studios)

Figure 4.21

Figure 4.21Manley SLAM! This is a unique and extremely useful stereo mastering compressor that features both an electro-optical compressor (à la the Teletronix LA-2A) and a FET compressor (à la the Urie 1176). These can be accessed in series, or independently as required. Each compressor has a number of mastering-centric modes. Early models had a digital I/O option that featured mastering quality Anagram Quantum ADDA converters (discontinued in 2009). The output gain has a very even un-hyped sound. It is a clever mastering-centric box and I use it daily.

Source: (courtesy Capitol Mastering)

Figure 4.22

Figure 4.22The Crane Song STC-8 is a discrete Class A FET compressor combined with a peak-limiter and an enhancement circuit for introducing analog warmth.

Source: (courtesy Crane Song)

Figure 4.23

Figure 4.23The famed Fairchild 670 variable-gain compressor was developed in the early 1950s by Rein Narma for The Fairchild Recording Equipment Corporation. It was used on many Beatles recordings by Sir George Martin, and was a mainstay in vinyl cutting rooms for decades.

Source: (courtesy Capitol Studios)

The most revered and iconic variable-gain compressor is the Fairchild 660 (mono) and 670 (stereo) (Figure 4.23) from the 1950s. The Fairchild has input, threshold, six different time constants (as attack/release times),10 and output available for user adjustment. They have a hallowed status in the pantheon of compressors due in large part to one Sir George Martin, who favored them on Beatles recordings.11 They utilized military-quality components and had vertical/lateral (M/S) functionality for lacquer disc cutting applications, so they were often installed in mastering studios before the advent of digital audio and the compact disc. They gained favor among newer companies seeking to create a reliable modern version, with the Manley Variable-Mu™ compressor (Figure 4.25) becoming a mastering studio staple. Other excellent modern examples of this topology are the Undertone Audio Unfairchild (Figure 4.24), Pendulum 6386/ES-8, Magic Death Eye Mastering Compressor (Figure 4.26), Thermionic Culture Phoenix, and Capitol Mastering CM5511 (Figure 4.27).12

VCA compressors utilize a voltage-controlled amplifier whose control voltage is derived from the audio input signal itself to effect the gain reduction. Classic examples are the dbx 160 (and permutations) (Figure 4.28) and dbx 165. These versatile compressor designs offer a great degree of user control for attack/release settings, making them excellent for stereo bus and mastering applications. Common VCA compressors suitable for mastering are the Alan Smart C2 (Figure 4.30), Neve 33609 (Figure 4.33), SSL G-Series (Figure 4.29), API 2500 (Figure 4.31), Overstayer 3722 SVC, and Vertigo VSC-3.

Figure 4.24

Figure 4.24The UnFairchild is a faithful modern-day recreation by UnderTone Audio (UTA).

Source: (courtesy UTA)

Figure 4.25

Figure 4.25The Manley Variable-Mu™ Stereo Compressor/Limiter is a mastering studio staple that draws on the design of the Fairchild 670.

Source: (courtesy Capitol Mastering)

Figure 4.26

Figure 4.26The Magic Death Eye Stereo Compressor (designed and hand-built by Ian Sefchick) is a variable-gain compressor that boasts impressive details, such as a special EQ feature in the gain reduction circuit and hand wound transformers.

Source: (courtesy Magic Death Eye)

Figure 4.27

Figure 4.27The Capitol CM5511 Stereo Compressor. Only four of these variable-gain compressors were built in 2011 by Ian Sefchick, and are used regularly at Capitol Mastering.

Source: (courtesy Capitol Mastering)

Figure 4.28

Figure 4.28A pair of dbx 160 compressors (designed by David Blackmer in the early 1970s) that use a VCA gain reduction circuit.

Source: (courtesy Capitol Studios)

Figure 4.29

Figure 4.29The SSL G-Series Stereo Compressor is a famed buss compressor, and was also onboard SSL consoles of the era.

Source: (courtesy Capitol Studios)

Figure 4.30

Figure 4.30After working with SSL and designing the G-Series compressor pictured in Figure 4.29, Alan Smart released the C2 Stereo Compressor. This one is from my mastering studio at Capitol, and imparts a variety of useful tonal characteristics, depending on the setting. ‘Crush’ mode yields over-compression with a mid-range boost, and higher distortion.

Source: (courtesy Capitol Mastering)

Figure 4.31

Figure 4.31The API 2500 Stereo Compressor is a mainstay of bus compression.

Source: (courtesy Capitol Studios)

Figure 4.32

Figure 4.32The Neve 2254/E (released in 1969) uses a diode-bridge topology for its gain reduction circuit. They were included onboard Neve mix and mastering consoles of the era, and later often removed and rack-mounted with a power supply—as pictured here.

Source: (courtesy Capitol Studios)

Figure 4.33

Figure 4.33The Neve 33609 evolved from the 2254 and was released in 1985. It remains a favorite stereo buss compressor among mix and Mastering Engineers. It also uses a diode-bridge topology for gain reduction. This is an early incarnation with no Neve logo or front-plate power switch.

Source: (courtesy Capitol Studios)

Expanders/Gates

Expanders increase dynamic range in audio by either lowering level below the threshold (downward expansion) or increasing the level above the threshold (upward expansion). Downward expanders are more common and function like an ‘inverted’ compressor by lowering the ‘valleys’ in the program material. Expanders are used far less often than compressors in mastering, but can be helpful in working with an over-compressed mix. The extreme version of an expander is a gate. With a gate, very little or no audio is heard when the audio signal passes below the threshold. I mention gates here for their relationship to expanders, but please note they have no practical application in a mastering context.

Brickwall Limiter (BWL)

A brickwall limiter (BWL) is a digital look-ahead peak-limiter that allows for a dBFS setting beyond which no audio signal will pass. Look-ahead refers to the main signal being delayed and the side-chain analyzed so that the limiter can process program peaks with ultra-fast attack/release times and a ratio of infinity:1. The result is extreme level control with no output signal above a user-set ceiling. The settings are usually threshold (with auto-gain makeup in some designs) or input gain, digital output ceiling, and release or time constant settings. A BWL is placed in zone 3 just after the AD converter, and prevents over-levels at the RDAW, allowing for unhindered loudness maximization. The first readily available BWL was the Waves L1 software in the mid-1990s, then in 2000, the Waves L2 hardware (Figure 4.34) was released, quickly becoming ubiquitous in mastering studios. Other early hardware BWL examples are the t.c. electronic M6000 (Figure 4.35) and jünger loudness control devices.

Informed by the intense limiting of radio broadcast compressors, the BWL proved transformative to the mastering world, ushering in a new era of the loudness wars. Responsibly and carefully implemented, a BWL remains a valuable and effective tool for a Mastering Engineer to achieve reasonable target levels. However, over peak-limiting has very undesirable artifacts including ‘hashiness,’ inter-sample peak modulation, inverted dynamics (loud sections shrink, and quiet sections overtake), and an uncomfortable listening experience. The BWL democratized a critical component of mastering—loudness—but many mix and Mastering Engineers began submitting peak-limited mixes that were afflicted with the artifacts described previously. Indeed, initiatives such as Apple’s Mastered for iTunes and streaming platforms such as Spotify that post level specifications are a direct response the pitfalls of overly peak-limited music.

Figure 4.34

Figure 4.34The Waves L2 (hardware version) is a digital look-ahead peak-limiter released circa 2000 that evolved from the Waves L1 plug-in. It utilizes 48bit fixed-point DSP and allows for a dBFS ceiling to be set, over which no signal will pass. The L2 was widely embraced and ubiquitous in mastering studios (along with the hardware t.c. electronic M6000 Mastering Processor) and demarcated a significant point in time in loudness enhancement, management, and the fabled loudness wars.

Source: (courtesy Capitol Mastering)

Figure 4.35

Figure 4.35The t.c. electronic M6000 Icon Remote, which controls the mainframe M6000 Mastering Processor. The M6000 implements 48bit fixed-point DSP, and in addition to quality peak-limiting, offers vast processing options including EQ, multiband compression, and expansion in both stereo and Mid-Side.

Source: (author collection)

BWL Safety—User Tips

A little peak-limiting goes a long way, and 1–2dB is plenty. One approach I regularly use is to begin with the BWL in bypass, and go for a good genre-appropriate sound and gain structure using the PBDAW and analog chain (zone 1 and zone 2), carefully avoiding over-levels at the RDAW. At this point, you should have a VU level of about +8dBu. Now engage the BWL and make your way to around +10dBu, which may be enough; otherwise, distribute additional gain at appropriate points in the mastering chain (before the AD converter) for a target level of around +12dBu. If the BWL is a plug-in, you will capture the relatively dynamic setting of +8dBu at your RDAW, then add a plug-in peak-limiter in the RDAW (zone 3). Using a BWL effectively requires judgment, care, and vigilance—the described process may require subtle macro-dynamic adjustments at the PBAW to preserve song dynamics. Note that with kick drum/bass-heavy genres such as rap or dance music, the kick will naturally swing beyond the VU meter levels indicated (Figures 4.364.38).

Figure 4.36

Figure 4.36The Voxengo Elephant is a brickwall limiter plug-in which offers a wide variety of coloration options and customizable parameters.

Source: (courtesy Voxengo)

Figure 4.37

Figure 4.37The FabFilter Pro-L2 is another feature-rich brickwall limiter plug-in worth auditioning and implementing for mastering. It includes extensive loudness metering.

Source: (courtesy FabFilter)

Figure 4.38

Figure 4.38The DMG Limitless brickwall limiter plug-in implements multiband dual-stage processing that separates dynamics and transients, and generates extremely smooth gain reduction curves.

Source: (courtesy DMG)

Digital Audio Workstations (DAWs)

A DAW is the computer-hosted software that allows for the recording, editing, processing, and delivery of the final audio master. For professional mastering, I recommend a two-DAW setup for the following reasons: the playback DAW (PBDAW) remains dedicated to playback, preliminary processing, and source level adjustments; the record DAW (RDAW) only captures the processed audio and renders/generates all required master formats; it allows for different playback and record sampling frequencies/bit depths; and it keeps your files organized as flat and mastered on separate drives. My PBDAW is Avid Pro Tools on a Macintosh, and my RDAW is Steinberg WaveLab on a PC. Some Mastering Engineers are devoted to Macintosh computers, others to PCs, or a combination of the two. I have also experimented with a two-PC system, and for that I recommend the excellent PC Audio Labs Rok Box custom PCs configured for audio work. The most common mastering-specific DAWs are Steinberg WaveLab, SADiE, Sonic Soundblade, Magix Sequoia, and Pyramix.

I realize many engineers use a single-DAW setup, especially those newly developing their mastering approaches. In a single-DAW setup, however, usually both playback and capture functions occur in Pro Tools, which is recording/mix rather than mastering-dedicated software. Additionally, playback and record sampling frequencies must be the same, often requiring sample rate conversion (SRC) of the source audio to meet any high-resolution requests. Also, final masters must be assembled in other software that can create a DDP or PMCD Master. I acknowledge it as a viable option, but don’t prefer it other than for an in-the-box mastering approach (see Chapter 16—In-the-Box Considerations).

Conclusion

Proficiency with the fundamental tools described in this chapter provides a solid foundation of skills for a Mastering Engineer. Well-modeled plug-ins represent a simulacrum of the coveted and expensive hardware equipment described in this chapter at a fraction of the cost, and often with expanded functionality. Despite this, desirable aspects of analog processing are hard to replicate in the computer. Conduct your own research and testing regarding the usefulness and fidelity of ADs, DAs, EQs, compressors, and BWLs. Perform A/B listening tests by changing just one piece in the mastering chain and carefully compare. This will help develop your opinion about what sounds better and why. Anyone can buy audio equipment, but informed and accurate knowledge of the tools help define a great Mastering Engineer.

Exercises

  1. What three devices make up The Primary Colors of Mastering? What specific function does each device provide the Mastering Engineer?
  2. If an all-analog signal path is selected, what two additional devices are required in the mastering chain (assuming a digital source file)?
  3. Define each measurement represented in the relationship equation: −14dBFS = 1.23V = 0VU = +4dBu. Which measurements are digital, and which are analog?
..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
3.233.232.160