Professional Analog Mastering.
Professional Analog Mastering.
226K+ subscribers
Mixing drums is a lot easier than it seems. Following the steps I’m showing here makes the process much more rewarding and less time-consuming.
Our session has 16 tracks - shorted and looped for the sake of hearing the before and after of effects.
It includes 2 overheads, 2 room mics, a crush mic, which is a heavily distorted mono recording, 3 mics on the kick, snare top and bottom, top and bottom mic for both toms, then a high hat and ride.
These were recorded and not sampled, but the concepts I’m showing can also be applied to sampled drums.
So, here’s how we improve the phase relations between the multi-tracks.
First, put a correlation meter on the stereo output - a stock plugin will work fine.
Next, solo the overheads. Pan the left overhead to the left and the right overhead to the right.
On the right overhead, insert a phase invert plugin - your DAW should have a stock plugin for this. Then, play the soloed overheads and observe the correlation meter.
The closer the meter is to +1, the better. The closer it is to -1, the more phase cancellations we have occurring between these 2 signals.
Notice that when I invert the phase of the right overhead, the correlation and measured phase relationships become worse. This indicates to me that these 2 signals have the best phase relationship when the right overhead’s phase is NOT inverted.
Since I won’t be inverting this track’s phase, I can take the plugin off and move on to the next track. I’ll keep the first 2 tracks soloed, while now also soloing the 3rd track.
The right overhead will join the left overhead by being panned to the left, and the newly soloed track will be panned to the right.
Again, we play the soloed tracks, and observe the correlation meter as we invert the phase of the newly soloed track. In this instance, inverting the phase is not helping - so again, I’ll remove the plugin, pan it to the left, solo the next track, pan it to the right, and repeat this process for all tracks.
The kick's boundary mic is the first track I’m noticing improves from being inverted. Notice that when I invert its phase, the correlation improves.
In the end, the kick boundary, the kick-in, the snare top, the high hat, and the ride tracks needed to be inverted to improve the correlation.
Of course, I’ll keep those 5 tracks inverted, return everything to the center, and then achieve some general panning and levels.
Let’s listen to the before and after - the levels will be identical, and the phase of those 5 tracks will be inverted. Let me know which one you think is inverted and which is the original in the comments.
Watch the video to learn more >
The traditional way to do this is to use a gate, but I prefer to edit the track.
If you use a gate to isolate the relevant signal and reduce track bleed, the plugin’s downward expansion and general behavior will alter the track's timbre.
However, editing will avoid this. To use a gate, simply insert your gate of choice and begin adjusting the threshold until the attenuation sounds natural. For drums, a hard knee setting works well on kick, snare, high hat, and toms, while more gradual gating works better on overheads, room mics, and cymbals; however, you may want to avoid gating these signals altogether.
The method I prefer includes opening up the editing window and while affecting the track - not the file (bear in mind that’s important if you don’t want to alter your audio permanently), and under functions, select remove silence.
Adjust the threshold until the relevant parts of the signal are isolated.
Notice that I’m being aggressive with the kick, snare, and toms, but am using a lower threshold for the overheads, rooms, and other similar tracks. For some tracks it may be best to just leave them as is.
Let’s listen to the original drums, then the edited ones, and notice how the edited drums still sound natural, just a little more controlled.
Watch the video to learn more >
Starting with the overheads, I will use identical settings for both. Since we’re dealing with a multi-tracked instrument, any changes to one signal will interact with another.
For this reason, I’ve used this same very gradual low-shelf and high-shelf setting on the overheads, the room mics, the crush mic, the high hat, and the ride.
Then, I’ll use identical EQ settings for each group, so to speak. So, the 3 kick mics all get the same EQ. It is the same for the snare mics, the high tom, and the floor tom.
The 3 kick mics have a low shelf boost of around 80Hz, a dip of around 400Hz to reduce knocking, and a boost of around 2kHz to amplify the beater hitting the skin.
The 2 snare mics have a 6dB/octave slope high pass to below the fundamental, which reduces unmusical lows, a subtle bell on the fundamental to increase the note and tone, a dip to 400Hz to decrease any knocking, and a moderate bell on a little under 5kHz to increase the snap and clarity.
The high tom mics have unmusical lows cut, the fundamental boosted, muddy frequencies reduced, that kind of basketball dribbling sound around 900Hz reduced, and some clarity added around 4.5kHz.
Similarly, the floor tom has its fundamental boosted, muddy frequencies reduced around 250Hz, that same 900Hz tone reduced, and some clarifying high frequencies boosted, again around 4.5kHz.
Let’s listen to the before-and-after. I won’t tell you which is which, but let me know in the comments which one has and which hasn’t been equalized.
Watch the video to learn more >
Compression can get complicated quickly, but I use this incredibly simple method that sounds great.
I’ll stick to the 1176, preferably REV A, with the blue faceplate and achieve 5-7dB of attenuation on every track, with a few exceptions that I’ll mention in a moment.
Here’s where I lose a lot of people, though; I always use the fastest attack and release settings. You can yell at me in the comments, saying I’m killing the transients if that’s what you want to do, but I’ve already converted a few very stubborn engineers into using this method.
Conventional wisdom says to use a long attack to let the transient through, but when we cut into the transient, we distort it, amplifying the highs.
I think a lot of engineers simply flip between fast and slow attacks and say, oh, slow sounds better - but I’m guessing that’s because they’re not compensating for the up to 5dB gain increase that a slow attack introduces.
When you compensate for the gain change, you’ll hear how much better this distortion sounds than simply letting the transient through.
I’ll use this method for everything except the crush mic since it’s already smashed, the hi-hat and ride, which just brings up a lot of ambiance, and I’ll get about 3dB of attenuation on the room mics.
Last thing to note, I’ll start with a 4:1 setting for most signals but move up to an 8:1 if I want additional distortion.
Let’s listen to this compression being introduced. I’ll show you one track that uses the slow attack setting and one that uses the fast attack setting - both compensated to have their peak levels as close as possible.
I won’t tell you which is which, but let me know which one sounds better to you.
Watch the video to learn more >
I know what you’re thinking - he just said this compression method doesn’t squash transients, yet here we are introducing transient expanders.
But honestly, transient expanders just sound good and are needed to get a snappy sound no matter what you do before or after.
So, I'm adding mild expansion on the kick, snare, and toms. I’m using this punctuate plugin, but I’m turning off a lot of what’s unique about this plugin to keep it simple, so use whatever expander you have.
All you have to do is use a low amount and dial in the time to encapsulate the transient best. Let’s take a listen and notice how we’re starting to get both a full and snappy sound.
Watch the video to learn more >
The snare, toms, and overall sound will quickly wash out if you don’t duck the reverb on drum tracks. The same thing happens if you reverberate tracks that simply don’t need it.
So before I show you some reverb settings, avoid adding reverb to room mics (since that would kind of defeat their purpose) - and any purposefully distorted tracks like the crush mic.
Everything else can have very intentionally dialed in reverb.
On the overheads, I recommend a high-density low-frequency reverb - then use a ducker to attenuate the start of the reflections and let the transient pass through unaffected.
The longer the release time, the dryer the sound - so this will take some dialing in while listening to the full mix to get a good idea of the interaction between everything.
On the kick, I recommend a kick chamber preset - something with a short RT60 and a good amount of low frequencies. Again, duck the hit and time the release to the song or to what simply sounds best.
On the snare, try a snare plate and adjust the decay time until it fits the style and genre - again, duck the hit, this time more aggressively, before blending in the amount to suit the track.
Toms sound great with dense reflections, like those found in an echo chamber, but with an adjusted reverb time of 1 second. The same is true for the floor tom, but with a reverb time of about 1.5 seconds. Again, we’re ducking the hit and then blending in the effect.
For the high hat and the ride, no ducking is needed - their dynamics are so intertwined with the full instrument that trying to control them is just an exercise in futility.
But, some airy reverb dialed in while in the context of the full mix will give these tracks an actual purpose.
One important thing to note, and it almost goes without saying, is that you should not over-reverberate unless you really want that type of sound.
In the last step, we’re adding some bus processing that’ll really amplify quieter details, reverb included, so use less reverb than what you’ll want in the long run.
Let’s listen to the reverb being enabled and notice how it subtly fills the space and keeps the transient detail intact.
Watch the video to learn more >
We're almost done. We just need to create a drum bus by changing all of the outputs of our tracks to a bus.
On the corresponding auxiliary track, we’ll introduce a plugin that maximizes the signal. My personal favorite is PSP Vintage Warmer 2.
Increasing the knee, the processor automatically amplifies the track, introduces soft-clipping, and adds harmonics.
Then, I’ll do some subtle tone shaping to emphasize the kick’s fundamental and the snare’s clarity.
I’ll switch the processing to multi to have the processing treat each range independently and then go into the settings to adjust the amount of saturation on each band and quicken the release times for the highs and mids.
If you don’t have this plugin, I also recommend the Oxford Limiter, Saturn 2, or the Omnipressor. But of course, substitute with whatever plugin you trust to affect the full instrument.
Then, I’ll add a M/S EQ to shape the sound subtly. I’ll make some of the lows mono with a side-image HP filter and then widen the mids and highs.
On the mid-image, I’ll balance the lows, reduce some muddiness, and increase the clarity a little more.
For the sake of transparency, I also added the plugin Fresh Air at the end of the Boundary Kick, the Snare Top, and the High Tom tracks to add some air and brighten the reverbr.
Let’s listen one more time to the drums as I introduce the Vinateg Warmer and then the final EQ.