For this chain, I want to get it to a loud LUFS but retain a clean sound - or in others words, one without significant distortion, pumping, aliasing, artifacts, or anything that doesn’t augment the quality of the track.
The primary strategy and what I’d recommend is starting with processing that either attenuated unwanted frequencies or prevents unwanted distortion.
That’s why the first half of the chain, made up of phase alignment, subtractive EQ, resonance reduction, and compression, all either attenuate or prevent distortion.
Then the second half made up of saturation, additive EQ, exciting, clipping, and limiting, allow me to add in what I want more of.
Let’s take a quick listen to the full before and after.
Before I add processing, I like to ensure that the phase of the track isn’t offset to either the positive or the negative value. If for example, it’s too oriented toward the positive side, some of my processors may misread the value of the peaks.
As you imagine this could affect when compression starts, clipping, saturation, or anything in which a threshold comes into play.
For this process, I import the track into Izotope RX, and then select the phase module, then suggest. This will show me how to offset the phase, at which point I can export the now phase-aligned track, and import it back into my session.
To illustrate this idea, let’s take a look at 2 tracks being compressed, one with an offset phase, and one that’s in-line, and notice how the compressor is attenuating more of the offset phase track.
The first insert in my master chain will be a Mid-side EQ - I’ll use this FabFilter Pro-Q 3 but a good free alternative is MEqualizer by Melda Audio.
I’ll create a band, change its placement to the side image, change the filter type to high pass, and then listen intently as I move the center frequency around.
The goal is to attenuate either right above or right below the fundamental frequency - making everything below the filter mono.
If I want a more centered sound, I’ll place the centered frequency above the fundamental - for a more impressive low end that spreads into the stereo field, cut below the fundamental. And if you’re mastering classical music or jazz, typically avoid this filter.
You may be wondering about the mid-image - some engineers like to attenuate up to 20Hz. to increase headroom by reducing inaudible signal - but to do this you’d need a higher dB/octave slope, which will affect the phase, in turn, amplify the lows above the cutoff.
So let’s listen to just this side image filter, and notice how the lows become centered by being isolated to the mid image.
Building on the processing from last chapter, I will use this same EQ to attenuate frequencies I want less of, and that I don’t want to be amplified by future processing. The frequencies you pick will depend on the track, so you’ll have to use your ears, but I’ve found that attenuating 250Hz. by .5 to 1dB will help the highs cut through.
This powerful frequency often masks higher ones, so attenuating it slightly reduces its impact on roughly 2.5 to 5kHz, or where we get a lot of a master’s clarity.
From here, I’m going to subtly attenuate any other frequencies I want - usually I’ll find a small section in the mids, and something near the vocal sibilance, but again use your ears and try not to attenuate more than 1dB.
Let’s listen to how this subtle attenuation slightly clarifies the track.
For this next step, I’m going to insert this Soothe 2 plugin - a less expensive alternative is Smooth Operator by Baby Audio or Track Spacer by WavesFactory, but I’ve found that this one works best for mastering.
It’s an intelligent EQ that uses multiple compressors to attenuate the frequency spectrum - in the middle, we can emphasize frequencies into these compressors, in turn controlling how much a particular range is attenuated.
Personally, I like to start with this preset - balance to the Grammys, which separates the signal into mud and side and attenuates them separately. Then I’ll listen closely and adjust the emphasis EQ until it sounds balanced.
Just using the preset will put you on the right track, but it is better to tailor it to your specific mix.
Let’s listen and notice how although the attenuation is subtle, there’s a noticeable difference in the clarity and balance of the track.
High-frequency compression often goes by the name de-essing, since this range of frequencies typically contains sibilance or ess sounds from vocal performance.
I recommend this Weiss de-esser since it’s really transparent, but a good free alternative is this T-De-Esser by Techivation. Whichever de-esser you use just be sure to use a split band setting, not a full band.
I like to use the high shelf filter type, then go into the settings, and set a slightly longer release and quicker attack to capture the transient quicker and smooth it out with the longer release. Then I’ll use a little makeup gain - now the highs will still be present but more dynamically controlled and with a slightly smoother sound.
Let’s take a listen.
You may not want to use full spectrum compression on your master, but mid-side compression is a good way to dynamically and naturally widen the stereo image. If I was to increase the stereo width with something like an Izotope imager, I would get artifacts and phase cancellation, which I want to avoid.
Instead, I use a mid-side compressor to attenuate the mid image more often and to a greater extent than the side image. This should happen with most mid-side compression settings since the mid is often more powerful than the side, and more likely to trigger the compressor.
Subsequently, whenever the mid image is compressed and the side is not, the level of the side image is increased relative to the mids, making the track sound wider.
What I like about this method, especially when trying to keep a master sounding clean, is that the effect is natural due to it being in time with the music.
Let’s take a listen, and notice how it gives the track a nice dynamic that’s enjoyable to listen to.
Saturation is a great way to make a master sound full - and if used thoughtfully we can avoid unwanted artifacts. I’ll use this Saturn 2 saturator, but a good free alternative is GSat+ by TBProAudio.
With it, I’ll introduce frequency-specific saturation - 4 bands usually work well. This way I can separate the track into my kick and bass, my low mids, the high mids where the vocal’s clarity is, and then the highs.
I found warm tube settings or anything that adds an even-ordered harmonic works well on the lows since it creates a full sound. Quick side note: To check what your saturator is adding to your signal, run a sine wave through it and measure the harmonics with an EQ or frequency analyzer.
Tape settings usually taper off some of the highs, while clean tube settings typically add higher frequencies and slightly increase dynamics. One more side note you can use the free plugin EQ Curve Analyzer to see if your saturator also incorporates some behind-the-scenes EQ.
Do what sounds best, but for the high frequencies I’ll usually use a tape setting and use it subtly - the reason being, if I generate harmonics on the high frequencies, I’ll like create one or more that go above the highest supported frequency, resulting in holdback or aliasing distortion.
To be sure I’m minimizing the presence of this aliasing distortion, which has an unpleasant sound, I’ll also introduce oversampling - which raises the highest supported frequency and uses a low-pass filter to attenuate any harmonics folding back down the frequency spectrum.
Let’s listen to how saturation fills the spectrum, and notice how tailoring saturation types to specific frequency ranges helps shape the sound.
Next, I’m going to insert another EQ - unlike chapter 4, I’ll amplify frequencies I want more of instead of attenuating ones that are causing issues. I like to use mid-side EQ and amplify a little around 300Hz on the side image - this mimics the depth function of the Neve portico, and adds some nice stereo width on frequencies that don’t need to be centered.
Then I’ll amplify the highs of the side image with a subtle shelf. On the mids, I like to amplify the kick’s fundamental, which I’ll have to look to find - and then amplify the clarifying frequencies of the vocals, somewhere between 2.5 and 5kHz.
This has been a good starting point for me, but I adjust as needed, and I’d recommend you do the same and use your ears to find what augments the track the most.
Let’s take a listen to the EQ being enabled.
Personally, I really like this Fresh Air plugin by Slate digital - it’s a free Air EQ, and exciter, meaning it amplifies the highest frequencies like an EQ and introduces harmonics to the high frequencies, causing a bright and detailed sound.
As we discuss in chapter 8, adding harmonics to high frequencies can be problematic due to aliasing distortion - Additionally, you’ll notice this plugin doesn’t offer oversampling to lessen the effect of aliasing.
So, I like to use this Metaplugin by DDMF to insert the plugin, and then include oversampling via this plugin wrapper. Then I’ll very subtly amplify the high frequencies.
Let’s take a listen and notice how the highs become brighter, and more detailed, and how we don’t have any unwanted artifacts.
Although counterintuitive, a clipper will help a master sound cleaner in the long run - whenever a transient crosses the threshold, usually, either the kick or snare, the clipper both attenuates the peak and adds some white noise. Although this is technically added distortion, the white noise makes the attenuated transients sound brighter.
I’ll use this Saturate plugin by Newfangled Audio to introduce hard clipping, but a great free alternative is FreeClip by Venn Audio.
As you could imagine, adding these high frequencies could cause aliasing, so I like to add this to the same plugin wrapper that I used last chapter and route it after the exciter.
Let’s listen and notice that although we’re adding white noise, the track’s transients still sound present.
For the last 2 chapters I’m going to break up my limiting into 2 stages - this way I can combine different limiting algorithms to shape the sound, and no one limiter is having to work too hard, in turn avoiding an over-compressed sound.
I personally really enjoy this Sonnox Oxford Limiter - and for a free alternative, I’d recommend Limiter No. 6 by Vladislov Goncharov.
I’ll turn off safe mode and auto gain, and increase the enhance function - this brings up quieter details, reducing the dynamic range but from the noise floor up, instead of the peaks down. This makes the master sound fuller and more impressive while avoiding reducing the impact of my transients.
I might engage the limiter, but I’ll do so by no more than 2dB.
Let’s listen to how the track sounds fuller, and louder, and notice how only a little attenuation is needed.
Last up, I’m going to insert a second limiter. For this stage, there are some things I want to avoid. I’ll be sure to turn off true-peak limiting, since this will reduce the impact of the transients, making the track sound less dynamic. Additionally, I’ll turn off or reduce any lookahead, since this will capture the transients more quickly, again resulting in less dynamic sound.
Interestingly, these 2 functions make a master cleaner from a technical viewpoint, by reducing inter-sample peaking and distortion to transients respectively, but when listened to, the sound is less transparent.
I also like to delink my channel’s detection - when linked the attenuation occurs to both the left and right whenever one channel crosses the threshold, but this way, I ultimately cause less attenuation, and add some nice variance between the left and right, making the track more dynamic.
Lastly, if it’s available, I recommend using a dynamic algorithm. Granted, find the algorithm you think sounds best, but I’ve found that this one in particular gives the track a more open feel while still getting the final LUFS to a loud level.
Let’s take a slightly extended listen to the full A B, with the peaks normalized to make them easier to compare.