Mastering was once the exclusive realm of specialists in lab coats laboring behind locked doors over mysterious and expensive equipment. Cut to the present, and now anyone and everyone with a DAW can do it—but it’s still a more specialized practice than many recording and mixing activities, and often does require a more delicate touch. Modern mastering tools are powerful, but the careless or improper application of them can easily do more harm than good to a mix. With that in mind, here are 7 things to watch out for when engaging in the black art of mastering...
One of the first things a producer/engineer/mixer will often hear when jumping into mastering for the first time is “Don’t!—let a ‘real’ mastering engineer handle it”. The people who offer this suggestion usually don’t mean to impugn the skills of the engineer in question—what they’re trying to say is that any project will benefit from a fresh perspective and a fresh set of ears at that critical final stage. No matter how skilled and experienced an engineer may be, after listening to a song hundreds of times, often focusing on the minutiae of the arrangement, it can be hard to step back and hear the music as a new listener hearing it for the first time would—what jumps out on that important first listen, what are the hooks that draw someone in, and what subtle flaws might be distracting, consciously or subliminally.
After countless playbacks, anyone would be likely to gloss over problems due to over-familiarity, or miss out on the opportunity to make more of a potential hook. A mastering person who’s really hearing the material for the first time will bring his/her fresh ears (and years of experience focusing on the “big picture”) to bear, and will stand a much better chance of zeroing in on those subtle aspects of the mix that need to be addressed for the best possible results.
Now, that’s not to say that you should never self-master a project you’ve also recorded and/or mixed—it’s often a logistical or financial necessity—but even if you do, you should be sure to have a trusted (and ideally experienced) listener or two to give the material a once-over. They can provide that fresh perspective, pointing up areas that you might want to pay extra attention to when you sit down to master it—those observations can sometimes make a world of difference.
The main tools used in mastering—EQ, Compression, Limiting—are the same kinds of processing used in mixing, but the application usually needs to be more subtle.
The choices you might make when EQing individual tracks, or compressing a particular part—the degree of tonal change or the amount and obviousness of gain riding—would typically be way too much when applied to the entire contents of a busy arrangement. EQ tweaks of 5–6 dB (common in mixing) are more likely to become very small adjustments of 1 or 2 dB—even a half dB or so—in mastering. That same presence boost that sounds so good on, say, the snare, would, if applied to the entire mix, result in a peaky, unnatural sound. The same amount of heavy compression that helps give a vocal added punch would almost certainly have a detrimental effect on background instruments if applied to the entire arrangement.
The processing in mastering needs to be applied with a much gentler hand—EQ tweaks are typically measured in single dBs and half-dBs; compression Ratios are more likely to be in the range of around 1.5:1–2:1, with lower Thresholds, to more subtly apply a gentle squeeze across the entire dynamic range. Mixers will want to shake off the usual approaches they’d bring to the individual components of an arrangement, and shift gears into that more delicate application of processing that will be suitable for treating all the elements of a mix at once without overdoing it.
The midrange is where most of the critical fundamentals and harmonics of the majority of musical instruments and voices lie. A good mix is careful to treat the all-important midrange area carefully, making adjustments at slightly different frequencies for different tracks/instruments, to keep from having too much energy either build up or be sucked out in the same ranges, giving the overall mix an overly hollow or boxy quality. But one careless application of midrange EQ at the mastering stage can easily undo all that careful work.
A lot of times, EQing in mastering may focus more on the extremes of the spectrum—high treble, low bass, getting the overall balance to sound good. Many modern (mastered) mixes sound like there’s a bit of a “smile curve” applied—slight boosts in the bass and treble, with a corresponding broad, subtle dip in the midrange, often centered around 1 kHz. This can work fine, but a curve like that should not be just applied as a matter of routine. If you listen to a number of different mastered songs, you’ll notice that some are a lot more mid-heavy than others—these songs may sound a little less “sparkly” at first blush, but as you continue to listen, their midrange detail may reveal more of the harmonic richness of the various parts, and ultimately offer up a more satisfying immersion into the arrangement.
Most of the compressors used in mixing are standard single-band designs, that affect the entire frequency range at once, but it’s quite common in mastering to employ multiband compressors. A multiband compressor divides the frequency range into (typically) 3, 4, or 5 separate bands, and compresses each range independently (though with proper phase-coherence).
Multiband compressors are powerful tools—in the right hands! By dividing the range, it becomes possible to apply different degrees of processing to different elements of a complete mix, for more flexible and transparent processing. The classic example is the kick drum that triggers the compressor, unfortunately pulling down the level of the entire mix every time it hits, resulting in the classic compression flaw of “pumping” or “breathing”. A Multiband can apply appropriate gain reduction to just the low-frequency band, controlling it as needed but still leaving the mids and highs alone.
On a more subtle level, multiband compression can function as a kind of “dynamic EQ” (though there are also separate processors that do that in a different way). Certain elements of a mix can be compressed when they push too hard without affecting other components of the arrangement as much—for example, over-loud cymbal crashes can be tamed momentarily without dulling down the entire mix every time they come along. Applying slightly different settings to several bands of a multiband compressor can be a more subtle and dynamic way to achieve slight shifts in frequency balance that follow the contours of the arrangement, without having to resort to more elaborate techniques like automation of a static EQ.
Besides offering an opportunity to subtly enhance a mix, mastering is also the last chance to fix any problems with the mix. While adding that commercial “sheen” and pumping up the volume are probably the obvious things that novice mastering engineers first turn their attention to, it’s of equal importance to use the mastering session to give the mix that last once-over, and detect any technical flaws that might come back to bite you in the ass later.
These issues include things like balance and phase, for example. The mastering engineer is the one who needs to ensure that the mix is properly balanced between left and right. This means not only that center-panned elements (kick, snare, lead vocal) are properly centered and not shifted off more to one side or the other, but also that the mix itself is well-balanced. For example, ideally there shouldn’t ever be too many rhythm instruments all lumped together on one side of the mix, with a big empty space on the other, as sometimes can happen in certain parts of a song as parts panned to certain positions drop in and out. Or all the brightest elements of the mix—hi-hats, cymbals, acoustic guitars, bells—shouldn’t be on one side, with the other side a duller soup of dense lower midrange energy.
The mastering stage is also where the phase coherency of the mix should be checked. Too much out-of-phase information can negatively impact the listening experience, especially on some systems—phase correlation meters can help ensure that phase differences fall within acceptable norms.
If you listen to enough commercially mastered songs, you’ll clearly hear that there’s a wide range of variation in overall tone and dynamic balance from song to song—some tracks are bright, thin and punchy, while others are thick and smooth. But they all fall within a range that allows them to sound good when heard in succession on the radio or in a playlist. One of the responsibilities of the mastering engineer is to make sure that the songs being mastered also fall within that range. They can still vary considerably in overall tonal balance, but if a song falls too far outside the norm—way too bright or dull, or much too heavily compressed—that song will likely suffer (subliminally) in comparison to other commercial releases, and that can impact subconsciously on the listeners’ impression of not only the mix, but the music itself.
Experienced mastering engineers have a pretty good internal reference for this, bolstered by the consistent use of neutral speakers in a carefully tuned environment they know they can trust. For the rest of us, maintaining a set of reference tracks can be a big help insuring that the material we master will stand up well against the best out there. These reference tracks should be full-resolution (no MP3s/MP4s) commercially mixed and mastered songs in a genre appropriate to the work being mastered, with a fairly wide range of different tonal/dynamic balances. As the ears become more accustomed to mastering concerns, these may become less critical, but when taking on a project in a less-familiar genre, they’ll still be invaluable to help keep things real.
This last topic has probably gotten the most press of anything related to mastering, so I won’t belabor it, but an article on mastering mistakes wouldn’t be complete without at least a mention. As I’m sure we all know by now, thanks to digital look-ahead brickwall limiters, it’s become possible to crank the average level of a mix up louder than ever before, by cleanly limiting (removing!) the transient peaks, allowing average (sustained) levels (which determine how loud something actually sounds to us) to be raised much higher than otherwise.
But the tradeoff is a loss of snap and punch (the impact of drums and percussive instruments), which can quickly turn a mix into mush. Artists and labels are afraid that if their songs are lower in level than others, the lack of sheer volume will result in a negative impression (lack of excitement) for listeners, but the listening fatigue that can be induced by that excessive level can be just as damaging.
There’s a point beyond which, no matter how good the processor, the mix will begin to suffer some loss of punch, and that point is routinely exceeded nowadays, often reluctantly, by mastering engineers who know better but must defer to the demands of the guys who pay their bills (those labels, artists, and producers). Up to about 6 dB of this kind of transient reduction can be transparent, and another 1 or 2 dB on top of that can still sound acceptable, but many masters push even another 2–3 dB beyond that for even louder tracks. Though modern brickwall limiters can do this more cleanly than ever, musically there’s still a price to be paid.
Fortunately, there’s a bit of a backlash against excessive loudness lately, and the practice shows some signs of subsiding eventually. I’d suggest shooting for a loud level (to satisfy the customer), but stopping short of that last 2–3 dB of boost, while the tracks still have some punch. That would mean if a mix maxes out at around 0 dBFS, you could get away with up to 6–8 dB of boost with the brickwall limiter (Maximizer), but no more. That may be enough to satisfy everybody, and still be a good compromise between loudness and punch.
Mastering is a pretty deep topic, but hopefully these suggestions will be of some help to both budding mastering engineers and weekend warriors who are just looking for the best sound they can get.