Professional Mastering - Analog & Digital
Professional Mastering
251K+ subscribers
To keep things organized, we’re going to address these points 1 by 1 throughout the video, building on ideas as we go. These points are:
PhaseDynamicsHz. ResponseTemporal EffectsBus Processing
That's the same order we address these elements as when we mix drums.
Watch the video to learn more >
When working with a multitracked instrument, some elements will be in phase, meaning they add together, and others will be out of phase, meaning they cancel each other out.
So, our first goal when mixing drums is to maximize the amount of in-phase signal. You might think we need to establish the levels, which you can, but we’ll just get rough levels and save a more methodical approach for later.
To start, we need an anchor, so to speak, or a track or tracks that determine the correct timing.
What you pick is up to you, but I always choose the overheads - they contain all of the drum’s info, and often give an accurate depiction of the instrument as a whole.
With the overheads organized at the top of the drum tracks, I’ll find the first big transient. In this case, I’m going to pick this snare hit.
I’ll zoom in and, one by one, align the other tracks’ transients with the overheads.
So, I’ll find the same snare hit on the room mic’s recording, and align it with the overhead.
Zooming in further, notice how the peaks and troughs align. It won’t be perfect - different mics from different positions record different information, but getting these transients to line up is the first step to optimizing the phase relationships.
I’ll continue this for any other ambient recordings and the snare’s close mics.
Then, it’s time to align the kick. Again, using the overheads as the anchor, I’ll align the transients of close kick recordings. If the overheads don’t have enough of the kick to align the kick’s close mics, then the room mics, or another ambient full kit recording will work well - it’s not perfect, but the goal at this point is to improve the relationship.
On occasion, you’ll encounter completely out-of-phase tracks - notice that this crush mic is 180 degrees out of phase with the overheads, the room mics, and just about everything else.
We’re not going to realign this track, since it’s not a timing issue, but a phase rotation issue. Meaning the peaks from drums registered as troughs on this crush mic, and the troughs from the drums registered as peaks. The timing was the same, but the polarity was inverted.
To fix this, use a utility plugin to flip the phase - your DAW should have a stock plugin that does this. This same issue commonly occurs with the bottom snare mic - if so, invert the polarity.
We’ll continue this process for all drum tracks - note that even if it looked like we nudged a track’s timing by a lot, we’re so zoomed in that any change we made was less than 1ms.
I used to suggest a quicker method for improving the phase, but this method gets the best results, and once it’s done correctly, all subsequent mixing becomes a lot easier since the foundation is solid.
So, the transients are aligned, but you probably noticed all this other info after the transient that may or may not be aligned.
Watch the video to learn more >
When mixing a highly dynamic, multi-tracked instrument like drums, gating becomes a phase corrector.
I went into a lot of detail about this in our last video; I’ll link it here, but in short, our transients are aligned now, which is great. But the mic bleed is not.
The snare’s transient, as recorded from the snare’s close mic, may align with the overheads, but this same transient, as recorded from the kick’s close mic, does not align.
By gating these aspects and isolating the close mics to their relevant transients, we remove all of this interference, or at the very least, we greatly reduce it.
So all kick mics, all snare mics, all tom mics, the high hat mic, and any other close mic that was used should be gated.
I’m going to use this Acon Digital DeBleed Gate for this, using the kick model for the kicks, snare model for the snares, and so on - I found this gate works amazingly for low-frequency instruments like the kick and toms, whereas I chose a different gate for the snare in the end. Just like compressors or EQs, it helps to have different options for the occasion.
Now, this plugin uses a neural network to determine the instrument, which I’ll vary until it sounds like just the transient is getting through.
Don’t be afraid to remove a good amount of the ambience from these close mics. What we lose from cutting these out-of-phase, strange frequency response versions of the drum’s ambience, we can replace with the purposeful ambience we captured with our room mics.
So, I’ll continue this process and somewhat aggressively isolate the instruments.
The next step is optional, but it really improves the sound.
I’ll use a unique version of a gate - this Acon Remix and adjust some levels on the overheads, rooms, and crush mic - basically any recording that captured the full set.
Since the overhead’s kick is slightly out of phase with the room mics, and I aligned the kicks’ transients to the room mics in this example, I’ll pull down the kick slightly.
I’ll also boost the snare to reinforce that in-phase relationship.
I’ll attenuate some of the high hat as well, since I want to use my close mic as the primary source for it. To ensure a good phase relationship between the stereo pair, I’ll make sure I use identical settings on both the left and right overhead tracks.
I’ll continue this process for any other full drum capture mics, to highlight or attenuate what I want more or less of, respectively.
Now that I have everything aligned and gated, I can get an accurate depiction of the drum’s levels. So, now is a great time to determine the levels with faders, panning, and so on. Otherwise, you’ll set levels, cut out a big part of the recording through gating, and then have to re-adjust.
Let’s listen to the original drums, followed by the aligned and gated drums. Just a heads up, at this point, the drums may sound a little sterile, but notice how big a difference we’ve made already.
Watch the video to learn more >
The compressor you choose for your kick, snare, toms, etc, should be done with a lot of consideration.
If you want a sharp, snappy snare, don’t choose a compressor that clamps down on the transient and then holds on forever - you want something that responds quickly.
If you want a resonant kick, you want something that can hold on to the sustain, and then, through makeup gain, bring up those details.
When mixing drums, compression, and other dynamic processing isn’t just about a consistent performance, it’s about controlling the timbre and the frequency response by adjusting the relative levels of the ADSR and introducing varying amounts of distortion.
I’m going to use a 1176 emulation, as a personal preference - this one is free, it’s by Analog Obsession, so if you need a free one, this is a great option.
At this point, soloing really comes in handy - the changes we’re making are so small that you’ll need to hone in on the single instrument to clearly hear it. I know people say don’t solo because you need to hear it in context, but with everything aligned and in phase, we don’t have to be terribly concerned about losing a sound we achieve when everything plays at once.
Starting with this snare top recording, I’m making tiny changes to the attack and release until I hear the snap I want from the snare’s attack and decay, and some of the resonance I want from the sustain and release.
With a compressor like a 1176, we’re making changes to the behavior in literal microseconds, so we’ll need to listen intently.
I chose this compressor so that I can do the following:
Let a small portion of the transient through unaffected, and then have the compressor clamp down quickly and aggressively.
This way, the original hit is present - as the compressor cuts into the middle of the transient, it reshapes the waveform, causing distortion. The release is timed to cover the range from the middle of the transient until the next snare hit - by attenuating this region, I can apply more make-up gain to increase quiet details.
This same logic is used with the kicks and toms. I find the right point in the transient to cut - letting the initial hit come through while introducing distortion from the middle of the transient onward.
The compressed section is made dense, but then amplified with makeup gain.
Conversely, say you want a smooth, maybe vintage sound - pick a compressor that captures more of the transient, doesn’t cut into it aggressively, and then holds longer to include subsequent transients.
An optical compressor, vari mu, or VCA compressor will do this well - just adjust the attack and release until you’ve achieved the timbre you want. A compressor will almost always play a huge role in a drum set’s frequency response, again, by affecting the relative levels of the ADSR, so spend some extra time on this step.
At this point, if you want to shape, emphasize, or attenuate some frequencies you weren’t able to address with compression, now is a good time to do so.
Maybe some of the kick’s fundamental needs to come up, or it sounds a little boxy in the low mids so I dip that, or the snares' high-frequency snap could be accented, etc. Since we’ve compressed and used makeup gain to bring quieter details forward, it’s easier to both identify and control the perception of these elements with EQ.
Just be sure to use identical EQ settings for any stereo pair - for example, the drum overheads.
Let’s listen to the before and after of compression and EQ.
Watch the video to learn more >
This is a highly subjective part of mixing drums - you might want super dry drums, or drums with 80s style gated reverb, or maybe just a nice layer on the snare.
I don’t like to add much reverb, and instead prefer to rely on the room mics.
But there are a couple of elements that could use some extra, maybe stylistic reflections.
For the snare, I believe a send works better than inserting the effect on the track. So, I’ll send both the top and bottom snare tracks to an aux track and use a reverb that fits the track and genre. Generally, a plate reverb works great on snare - I recommend finding a reverb with ducking to let more of the original transient through, while the later reflections come through after.
If I find the right sound, I may send a little of the toms to this same reverb, but again, we’re talking about subjective decisions at this point.
Lastly, for reverb, I’ve found a moderate amount of short ambient reverb working incredibly well on the high hat track. With the short stereo variation that a stereo reverb provides, the high becomes less of a metallic thump and more of an airy and splashy sound that’s a lot more enjoyable to listen to.
Watch the video to learn more >
This is where a lot of the cohesion and fullness is accomplished. As you can see here, I have 6 processors - this may be too much for what you’re doing, but each step is playing an important role.
First, we have compression with a little makeup gain - it’s subtle, but by establishing a cohesive timing for the drum’s ADSR as a whole, it begins to tie everything together.
Next, I added a plugin I’m working on called Sage Velvet for some saturation. It uses a neural network recreation of a popular saturation circuit - so everything is very program dependent, but in short, it’s adding some mild to moderate harmonics to fill the sound, and I’m widening the over drum set slightly with this width function.
Next, I found I needed some frequency-specific stereo control, so with a mid-side EQ, I cut some of the side from the lows to center the kick, expanding a small amount of the high mids and the air.
I centered the kick’s fundamental slightly and centered some of the snare’s overtones with mid filters.
Then, I’m using NewFangled Audio’s transient expander and then their EQ to first subtly enhance the attack and decay of the instrument as a whole, and then, with the EQ, amplify or attenuate critical bandwidths.
With the preset called Bark Bands, the filters are centered on frequencies that don’t conflict with one another due to how our ears process and group sounds. I find this is a perfect way to reduce some masking and help some lost details cut through.
To finish the chain, we have this PSP vintage warmer. It’s limiting when needed, but I’m mainly using it for the upward processing that occurs with the knee function.
With it, all of the quieter details are pushed forward, creating this incredibly dense and detailed sound.
I’ll use the multiband option, and then in the settings window, I’ll reduce the attenuation time to the mids and highs to retain some detail, but keep it at the default level for the lows to avoid any distortion from quickly affecting the kick.
Since I’m bringing a lot of detail forward, the room’s ambient and the reverb are amplified significantly.
To account for this, I’ll then adjust the room mic’s level to dial in how much of this ambience I want.
Additionally, I routed the snare’s reverb to this bus, so the reverb aux track is now much easier to hear. So I’ll adjust this as needed as well.
With some final small changes, maybe small problems I didn’t notice before fixed here and there, these drums feel great.
Of course, other instrumentation will affect how these drums sound, but with a great foundation, I’ll likely only need to adjust a level here or there to work with the full mix.
Let’s listen to the drum bus processing enabled one-by-one. They all accomplish something and build up to the PSP processor, so that everything I want to bring forward with the knee function is there and ready to be amplified.