Chapter 14

Advanced Mastering Chain Tools and Techniques

This chapter explores advanced mastering methods. These techniques are more complex and nuanced, involve a greater margin of error, and require the confidence of execution that only experience provides. Used skillfully, a Mastering Engineer can achieve impressive, high-quality results. As always, it is critical to lean on the fundamental habit of referencing your flat mix religiously so that your adjustments represent a genre-appropriate optimization rather than a dramatic reinterpretation of the mix.

Side-Chain Compression (s.c.)

Most mastering compressors are equipped with an onboard s.c. function or external s.c. input (or both). This is so a ‘mult’ (duplicate) of the program material can be EQ’d and sent to the s.c. input, thus affecting the frequencies the compressor does or does not react to. Two popular examples are a high-pass filter, so that the kick drum does not cause excessive compression or pumping (the audible attack and release of the compressor)—often onboard in mastering compressors; and a de-esser, whereby a high-frequency boosted version of the song is sent to the s.c. so that the compressor reduces those same frequencies. Another s.c. option is sending a pre-compressed version of the audio to the s.c. input so that the compressor functions more gently on transients, even with fast attack times. Examples of quality mastering compressors with side-chain functionality are the Manley Variable-Mu™, Manley SLAM!™, Pendulum Audio OCL-2, Shadow Hills Mastering Compressor, Magic Death Eye Mastering Compressor, and Alan Smart C2.

Multiband Compression/Limiting (MBC/L)

This is a compressor or limiter that separates the audio spectrum into separate frequency bands that can be adjusted independently of each other. Imagine each band as its own independent compressor or limiter with its own threshold, ratio, attack, and release. Some companies even include harmonic saturation/distortion and also M/S functionality. A multiband dynamics processor can change the instrument and frequency balances significantly, so I’m careful not to overuse them and alter the mix excessively, imparting too drastic of an imprint on the source audio. A great mastering job is always relatable to the flat mix and should be an enhancement of what the recording and mix engineers sought to achieve. I would recommend MBC/L for mixes that are problematic, or for genres that require an extreme ‘hyped’ result—sometimes found in pop, dance, electronic dance music (EDM), electronica, or possibly some sub-genres of hip-hop/rap (Figure 14.1 Figure 14.2).

Figure 14.1

Figure 14.1The UAD Precision Multiband offers five spectral bands of dynamic range control. Separately choose compression, expansion, or gating for each of the five bands to manage everything from complex dynamics control to basic de-essing.

Figure 14.2

Figure 14.2The Weiss DS1-MK3 is a standalone digital hardware dynamics processor that handles de-essing (with low-pass, band-pass, and high-pass selectable compression band), compression, and limiting functions.

Serial Peak-Limiting

This technique involves the risky outcome of a loud result with very diminished dynamic range and works most effectively on sparse but punchy music such as rap/hip-hop, dance/electronica, or EDM. Serial peak-limiting involves the extremely bold move of setting up two peak-limiters in series either as plug-in instances, engines in the t.c. M6000 hardware, actual L2 hardware units, or a combination of these options. The peak-limiters are in zone 3, after the AD converter and each one generally adds approximately 1dB of gain below threshold. The output ceiling can be set to taste, but I generally use two different plug-ins with a ceiling of −0.3dB for the first peak limiter and a ceiling of −0.6dB for the second one. Listen very carefully to the effect of each limiter, and if the audio fidelity deteriorates, remove one or both.

Dynamic EQ

Although I primarily use dynamic EQ in mastering for zone 1 de-essing, it is worth exploring. Like MBC, for mastering applications, dynamic EQ may help with some surgical fixes on specific elements or for additional tonal balancing within mixes. A dynamic EQ combines the band-separated aspect of a multiband compressor with the linear changes provided by an EQ. As such, in addition to setting EQ filter parameters (frequency, boost/cut, shape, and Q), there are dynamics parameters such as attack, release, threshold, and even s.c. options. This means the selected EQ is interacting with the dynamics of the music and engaging as needed. There are several good plug-in examples available: brainworx dynEQ_V2, The Ozone 8 Dynamic EQ, and Hofa IQ EQ3.

Saturation

Saturation is a type of gentle distortion that can add some vitality and dimensionality to a mix that is lifeless or contained-sounding. Saturation is a byproduct of overloading the input or output of analog equipment like tape machines, tube amplifiers, or console microphone preamplifiers to create harmonic distortion. The increased input amplitude exceeds the capacity of the electronics or analog circuit to cleanly pass the signal. This can result in a richer and warmer sound, up to a point … beyond which the mix will sound overly distorted.

There are digital emulations of saturation as a feature in both hardware devices and plug-in software. For example, the t.c. electronic Finalizer includes a DRG (digital radiance generator) function allowing the user to dial-in emulated saturation, and the Crane Song HEDD-Quantum AD converter has options for tube and tape saturation emulation. In addition, the DMG Equilibrium plug-in has a ‘harmonics’ option for either a high-pass or low-pass filter that saturates the unfiltered (remaining) frequencies in the audio spectrum, and following I will review tape saturation plug-ins, as well. These are effective tools to have in your mastering arsenal, but be very careful with saturation as you are deviating further from the qualities of the flat mix.

Analog Tape Compression

One original reference point for saturation is tape compression, whereby the input signal exceeds the capacity of the metal oxide particles in the tape to cleanly reproduce the given amplitude. This results in the tape becoming saturated, sounding like a combination of both compression and subtle saturation. The standard analog playback and record tape machine alignment settings—repro level, record level, record bias,1 high-frequency and low-frequency adjustments—allowed recording engineers to align the machine to match tape manufacturer specifications, or to elevate the record gain adjustment and ‘hit the tape harder’ for a louder, punchier, and more saturated sound. With the advent of digital audio, this time-consuming and expensive analog process is available virtually in the form of various tape machine emulation plug-ins. Popular options are Slate Digital VTM, Avid Reel Tape, and UAD Magnetic Tape Bundle (which includes both Ampex and Studer virtual tape machine options). Of course, if you want to be a ‘mastering cool kid,’ you can get an actual analog tape machine.

If you have access to a good analog two-track machine, some quality recording tape and a bit of time, you can transfer a digital source file to analog tape before mastering to add warmth and dimensionality via tape compression. Keep in mind that analog two-track was the mix engineer’s format of choice for several decades. In my comparisons, the Ampex ATR 102 two-track machine sounds ‘quicker’ reproducing transients, making it great for rock with live drums and pop styles of music. The Studer A820 two-track machine—which handles tape superbly—seems to round out certain transients and sounds ‘slower’ on the reproduction of drum transients.2 The 15 inches per second (ips) tape speed shifts the machine’s ‘head bump’ (low-frequency response buildup on playback from the tape machine) lower and creates additional richness or warmth. Tape at 30ips is generally a bit more ‘glassy’ and extended sounding in the high frequencies, but still a great option for an extra analog element in the mastering process.3

My favorite configuration is Ampex 456 tape at a +6 record level alignment4 at 15ips on an Ampex ATR 102 two-track tape machine with a half-inch head stack. I then print three separate samples of the song—usually about one minute of a chorus—at three different levels (increasing in increments of 2dB) to play back and make a final determination of which level sounds best and presents the most appealing tape compression sound. Remember to not get seduced by the process and to always use your best judgment. This means that sometimes the digital file is still the clear winner.

Clipping

Another method to create a louder or more aggressive result is called clipping. The simplest way to do this is to increase the last gain stage before the AD converter thereby clipping the converter for a louder result accompanied by characteristics slightly harder than saturation. Telltale signs of clipping are squared-off peaks in a waveform. Another method is to have an analog clipper circuit in your signal path placed just before the AD converter. This allows you to push gain and achieve apparent volume, as well as protect the AD converter from producing unpleasant artifacts from overloading. These use a discrete component circuit of transistors, Zener diodes, or even LEDs in the feedback loop of an op-amp. The clipping can be designed to be either a softer or more rounded response at the +18dB point for a gradual or warmer sound, or harder and more angular, yielding a harsher more quickly distorting sound. Common clippers are the Prism Sound ‘Over-Killer’ barrels (a diode circuit that clips signal voltages above a +18dBu level) and the Clip +18 setting on the Manley SLAM! The +18 clipper is designed to function in a mastering system that is referenced to −14dBFS at 0 VU = 1.23 V = +4dBu (if −14 = +4dBu, then +18 will be at 0dBFS at the AD converter) (Figure 14.3).5

Figure 14.3

Figure 14.3The Magic Death Eye Evrenizer is a one-off (serial #0001 of 1) analog clipper that gradually switches between soft and hard clip settings.

Analog clippers will clip and eventually distort, and even pass signal above the clipping threshold. In contrast, digital BWLs will not pass signal above 0dBFS. The original Apogee AD series converters had a soft limit function to prevent over-levels near 0dBFS. At Capitol Studios in the mid- to late 1990s, the tech staff under Jeff Minnich6 built the fabled ‘black box,’ which the technical staff originally made for Mastering Engineer Wally Traugott to create hotter masters.7 It was a clipper that would remove signal above a certain threshold. As it was an analog device, if too much level was sent through the unit, it would pass over-levels, so the digital BWLs (such as the Waves L2 or the t.c. electronic M6000) provided a cleaner and more reliable option.

Parallel Processing (Compression and EQ)

Parallel processing refers to audio signal that is multed then processed—usually more aggressively than normal—and then blended in with the original stereo signal. You must always verify there is no delay or latency between your main program and the parallel program while doing this, or the result will be much worse than the flat mix.

As I consider parallel EQ as generally corrective, I perform it in zone 1 of my mastering system, meaning the PBDAW. This is also due to the incredible set of surgically adjustable parameters in many mastering EQ plug-ins. My approach for tonal balancing is to set up parallel EQ with a certain band of frequencies boosted and the remaining bands cut quite aggressively or high/low-pass filtered. If a mix is light in the low frequencies, for instance, I can blend in a low-boost stereo pair to compensate. Or, if the vocal is lifeless, I can blend in mid or high-mid frequencies until they just begin to affect the detail of the vocal. What you adjust affects everything in mastering, so consider M/S EQ for your parallel work to minimize the overall footprint (Figure 14.4).

For parallel compression, I prefer the sound of analog compressors and use the analog blend fader on my mastering console to add in detail from a compressor set for 4–8dB of gain reduction, which would be considered maniacal in a standard mastering context. Be careful with transformer or tube-designed compressors for parallel applications, as they may introduce a very slight delay that will smear the audio image. Many dedicated mastering consoles allow for parallel functionality with a stereo blend knob/fader such as the Dangerous Master and Liaison, Maselec MTC-1X, SPL DMC, and Crookwood Mastering Consoles.

Mid-Side Processing (M/S)

M/S Processing represents an important advanced mastering technique which I cover in detail in Chapter 15—Mid-Side—An Elixir of Mastering Hope. The methodology originates from the Blumlein stereo recording technique, and allows for mix elements in the middle of the stereo image to be processed separately from elements in the sides of the image.

Figure 14.4

Figure 14.4An ITB ProTools mastering session setup up with low and high-frequency parallel EQ blended in before being routed to a mastering chain on the master fader. This could also be an advanced PBDAW zone 1 setup before converting to analog for zone 2 and then capture in zone 3.

Hybrid Domain Mastering (Both Digital and Analog)

This represents less of a technique and more of an understanding of increased complexity in the mastering system. As opposed to either performing ITB mastering, an all-digital hardware chain, or a primarily analog chain, a hybrid domain system incorporates all of them. As outlined in Chapter 3—The Mastering Studio, I separate the mastering system into three distinct zones: zone 1 is the PBDAW and any digital processing before DA conversion, zone 2 is the analog equipment chain in the mastering system, and zone 3 is the region after AD conversion where any digital hardware processing (DSP or peak-limiting) and then any RDAW adjustments occur. Incorporating aspects of both digital and analog processing represents the best-sounding fidelity for most mastering projects. There are certainly exceptions, as I will discuss in Chapter 16—In-the-Box Considerations. In my many tests and comparisons, The Eleven Qualities of Superb Audio Fidelity from Chapter 2—Listening Experience remain most enhanced in a hybrid domain system. Note this does not require excessive processing in each zone—at times simply setting up the gain structure to achieve the desired target levels is all that is necessary.

Line Amps and Transformers

Mastering Engineers occasionally seek to add line amps or transformers to their analog signal path for alternate gain staging, coloration, vitality, or light saturation. These devices are custom built by the Mastering Engineer or a trusted technician. Suffice it to say that all Mastering Engineers are on an enduring quest to set their work apart vis-à-vis custom equipment or modifications.

Stem Mastering

Stem mastering involves the mix engineer separating mix elements into separate stereo pairs. The simplest example is a vocal stem and an instruments stem that combine to make the final mix. Stem configurations can become increasingly complex; for example, using four stems (drum stem, bass stem, music stem, and a vocal stem) is also common. This approach between mixing and mastering allows for stunning results without having to compromise enhancements (the Mastering Engineer’s eternal dilemma). To illustrate, by boosting 4.8kHz, many mix elements are affected—a snare may be enhanced enough, but a guitar may be simultaneously compromised by sounding edgy. With stems, for example, the bass frequencies can be enhanced without affecting the vocal, the vocal can be brightened without affecting the cymbals, and the kick drum can be enhanced without making the bass guitar boomy.

Stem mastering also allows for analog summing of the final mix, as the stems are played out of the PBDAW, through the analog mastering chain, and recorded into the RDAW. However, this approach can increase margin of error if the mix engineer has balanced the full mix through a limiter, then generates stems through the same limiter setting. The lower level of each stem will not affect that limiter as intended, and instrument balances will be off. It is best to mix without excessive limiting if you will be generating stems for a Mastering Engineer. Although this is not a hard-and-fast rule, most stem mastering adjustments occur in zone 1 (the PBDAW), before routing to the stereo analog chain.

Intercuts (Mastering Moves by Song Section)

There are instances when a song section requires a different treatment than the rest of the song. Some examples are an instrument solo that is slightly buried in the mix and needs more high-mid EQ, a bridge that sounds better with more low frequencies accentuated, or a loud climactic ending where cymbals become overbearing and require EQ or de-essing. Although these changes can be attempted ‘on the fly’ in real time, it is best to go back and record the sections with different settings into the RDAW as an intercut and edit them into the original mastered pass of the song.

These surgical mastering moves can make a significant impact in improving the mix and final master. It is imperative to be extra careful with this type of editing and zoom in to verify that each seam of the edit(s) line up perfectly. If there are several intercuts in your project, consider rendering the file to permanently capture the edits and name in a new session (while saving the previous edit stage just in case) so they don’t inadvertently get moved or altered, causing alarming drop-outs or skips in the master. These edits and captures must be performed on the high-resolution file first captured in the RDAW, maintaining the original sampling frequency and bit depth (as mentioned, I regularly use 96kHz—32bit float), not the rendered Red Book .wav file (44.1kHz—16bit).

Adding Reverberation

Reverberation or reverb refers to the natural decay of a sound source in a room or hall as the sound waves bounce off of the walls, ceiling, and floor. Early reflections are the initial reflected sound waves. Reverb time is determined by the room volume and absorption. In the early days of analog tape, recording a musical performance was captured live to a mono or stereo tape machine, and if the recording studio was acoustically absorbent (dead), reverb could be added via a live echo chamber to achieve a more natural, acoustically blended sound. The final application of reverb was often left to the mastering session. Later, with the advent of multi-track recording and other reverb options (spring reverbs, reverb plates, digital reverb units), the addition of reverb remained the domain of the mix engineer.

At Capitol, there are eight live echo chambers8 underneath the parking lot. These chambers were reputedly designed by Les Paul, and constructed initially as a group of four when the building was built in 1956, with four more added soon thereafter (Figures 14.5 and 14.6). Marquee artists ranging from Frank Sinatra and Nat ‘King’ Cole to The Beach Boys, among many others, have utilized them on legendary recordings. I am able to access these chambers from my mastering studio, and on certain recordings—primarily singer-songwriter, solo instrument, or some sub-genres of rock—it can add a unique vitality and ambience. A common airspace on certain recordings imparts a natural and familiar listening experience. At any live musical performance, there is a single concert hall or venue that imparts its particular reverberation characteristics to every instrument and sound. Of course, there are myriad plug-in options for reverberation, as well, and UAD has recently released a plug-in version of the Capitol echo chambers. From a mixer’s perspective, adding reverb to a mix at the mastering session may seem problematic at best, but it does have a place in the Mastering Engineer’s arsenal of approaches.

Figure 14.5

Figure 14.5The original floor plan of the first four Capitol echo chambers built in 1956.

Source: (courtesy Journal of the Audio Engineering Society)

Figure 14.6

Figure 14.6Construction of the Capitol echo chambers.

Source: (courtesy Universal Audio/UAD)

Conclusion

These advanced mastering chain tools are best incorporated into the mastering system gradually, if The Primary Colors of Mastering are not providing the desired sonic result or impact. The Mastering Engineer may develop their approach so that two or more of these methods are implemented into the same mastering system. It is best to add one technique at a time in order to discover its strengths or weaknesses, and determine how it affects the fidelity of the mastering project.

Exercises

  1. Set up a mastering compressor with a side-chain input to pass audio with about 2dB of gain reduction. Try three approaches altering the side-chain input (de-essing, high-pass filter, compressed audio). Keep the compressor settings the same except the threshold to isolate the side-chain effect. Notate how the compressor function changes in each instance.
  2. Using only an MBC/L or a dynamic EQ plug-in, master a song. Compare the result with the flat mix. Notate the result(s), and explain if you consider it an improvement or not.
  3. Add a tape emulation plug-in to the same song and mastering chain used in Exercise 2. Compare the mastered files. Describe how the tape emulation plug-in changed the new master.
  4. Set up a simple analog mastering chain (source file–DA converter–EQ compressor–AD converter), then clip the AD converter by increasing the compressor output gain, and record the result. Compare to a recording of the same song and setup without clipping the AD converter. Notate differences between the two files.
  5. Master a song in ProTools (ITB) with only a parallel EQ setup and a peak-limiter. Describe how you arrived at the final settings.
  6. Master a song with a vocal using only a M/S EQ plug-in, a compressor and a peak-limiter. Enhance the vocal only then capture, enhance the kick drum only then capture, finally enhance the guitars/keyboards only, then capture. Play each of the three versions, then describe how you utilized M/S EQ to affect only specific areas of the mix.
..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
3.235.182.206