|March 31st, 2011, 09:46 AM||#1 (permalink)|
Join Date: Apr 2002
Location: Melbourne, Australia
Systematic Mixing Series #2: In Soviet Russia, Drums Slam You
The full version of this tutorial, along with many others is available in the Systematic Mixing Guide, available here: http://www.systematicproductions.com/mixing-guide.htm
Beware, if you start scrolling down now you may never stop! This one is massive. Never did I picture when I started writing that it would amount to so much, and at the end still leave so much unsaid. Drums truly are one of the most involved elements of any production, as is evidenced by how much there is to go over. Recording and processing them is a personal affair - driven by many factors - and it would be impossible to list every approach. So here we go for a broad approach, talking about common starting points and problem zones for standard rock and metal mixing. We also go over samples, their role in final drum sounds, and how to optimize their use.
Drums are one of my absolute favourite things to mix. They tend to form the very extremes of a mix, from the cymbals - which are generally the highest and airiest of elements - down to the kick, which in many cases is the lowest. As such, they become the framework of a good mix, and thus getting good drum balance is critical.
In the modern era, as many of you would be used to dealing solely with sampled drums, as there are those who deal with raw drumkits. The principles when working with both do vary slightly, but at it's heart the goal is always the same. You want to find a sound that suits the music.
For the purpose of this tutorial I'm going to focus largely on my own approach, and the one I feel creates the best balance of vibe and punch - augmenting real drum kits with samples. This means there will be a large focus on keeping as much of the raw drums as possible and/or desirable. The principles will be equally applicable, albeit less intensive, to those dealing solely with sampled drum kits.
Skimming over tracking
Drum recording could be a whole bible unto itself (and actually is, if Glenn Fricker has anything to say about it!) so we will only briefly touch on the ideas behind tracking drums, and instead spend the majority of our time talking about how to mix them into a record.
If your intention is to keep as much of the raw kit sound in your final mix as possible, it goes without saying that you need to get the fundamentals right. Unfortunately for you, drums are arguably one of the most difficult elements to get right across the board when tracking.
Here is a quick checklist of some of the things you need to get right in order to succeed in keeping a quality raw kit sound on your record:
-Very good player, who can hit consistently, and consistently hard.
-Very good drum kit, which sounds open and minimizes muddy resonances.
-Very good cymbals, which have minimal harsh resonances, and minimize their crossover with the vocal range (so, not big and low pitched!)
-Brand new heads, ideally switched after every song.
-A knowledge of how to tune those heads optimally (much harder and more creative than tuning guitars).
-Very good room. For drums, the room tone is crucial, as it becomes a significant part of the drum sound in our mix.
-Minimal spill in your close-mics.
-An array of great quality microphones.
-An array of great quality microphone preamps.
-Good quality, 16-channel AD/DA converter.
So - intimidated yet?
Recording quality drums costs money - a lot of it. That's just the reality of something which has so many point sources that need to be isolated and treated individually to each other on a record.
It's no wonder that more and more projects are skimping on drum tracking, and instead using samples to drive the record. Many are even completely circumventing real drum kits, and using entirely electronic drums.
Now, in many ways we can circumvent those above requirements, depending on the type of project we are working on, and the type of drum sound we are shooting for.
Modern metal has the advantage of currently sporting an aesthetic which favors very direct drum sounds, with minimal room and acoustic spill. This allows us to, in many ways, circumvent our need for a great room. In fact, if you are able to record drums in a decently dead room, whether it be small or otherwise, you can then isolate and process your drums with more precision, and simply apply reverb, or use room samples (we'll get to that) in order to create an artificial sense of space. This gives you a greater sense of control, and for many ultra-modern sounding metal productions is actually favorable over using a large, live drum room and acoustic ambiance.
Rock music, on the other hand, can be a little more reliant on natural acoustics. A quality drum room is desirable, even if simply to retain the tightness of the bottom end on room and overhead mic tracks, as they are often used to solidify and enlarge the size of the drum sound.
Getting to point A
There is a lot involved in preparing drums for mixing. The tracking and editing phases are usually substantial and very involved. A lot of an engineer's personality can be found in how they tackle drums. As much of this is so personal, I can only hope to give a rough outline of common procedures, and what you may or may not wish to do with drums.
For the purpose of this guide I'll assume that we at least have an acceptable raw drum sound to work with.
The very first thing you'll want to do after the drums on a project are all tracked is pull the DAW session up and start checking phase. Good phase is critical to raw drum sounds, and can be the difference to something which stays thin and washy, no matter how much you process, to something which is dense and solid.
There are two major ways to go about correcting phase after a recording.
The old-school way is to simply sit there, listen to the overheads, switching the close-mics in and out, and inverting the phase on the channels, listening to hear which sounds 'larger'. The dead-give away is the setting which gives the drums more body. The more body, generally the more in-phase the drum.
So that's fairly simple, right? Well, yes, but it's also somewhat limited. If your drums happen to be 90 degrees out of phase with the overheads, then inverting the phase will do nothing for you.
How do we get around this blind-spot? Well, that brings us to our second method for correcting phase, which fortunately modern DAWs have made very convenient and easy to do.
You simply bring up your overhead tracks, and zoom right in until you can see the transient hit of whichever drum you're correcting. Now pull up the close-mic track of that drum, and nudge it back in time until it roughly aligns with the overheads. Note, this won't be perfect, because nothing is perfectly in phase with drums - but simply getting it closer should give you a larger sound. Repeat this for all your drums, and you should find a sound that is immediately more deep and dimensional, and creates a much better starting point for you in the mix.
This part is a love/hate of modern metal engineers. In many ways this part of the process is where the magic truly happens, as commonly average performances are tightened up into something inhuman. It's worth understanding that the vast majority of drummers out there don't have what it takes to record a world-class quality performance on drums - in metal, with its athletic demands, this is even more the case. You have to sit back for and while and wonder about what sort of unique individual it takes to willingly want to sit down and bash drums all day. Some do it because they aren't adept at any other instruments, or because it was the easiest way, but others – a rare few – are truly gifted and invested in their percussion. With that in mind, cast yourself back and realize that 99% of your time will constitute of working with the first kind. A lot of the time you simply need to throw the kitchen sink at the drums (if you're not tempted to include the drummer in that metaphor too) just to make it sound like a passable performance. This is relatively normal when working on modern music in this day and age, so grit your teeth, lube up and get ready to bear with me.
There are multiple ways to edit drums, depending on the DAW you use, and once again, this could be a tutorial series unto itself, so we will only touch on the concepts, and less the techniques.
The first order of the day, generally, is to tighten the timing of the performance. This can be done a number of ways, from slip editing, to using a Beat Detective style tool, to using an Elastic Audio style tool.
My preferred method is the use of a Beat Detective style tool which is capable of specifying the strength of quantization. Slip editing tends to result in sounds that are a bit too rigid for my taste. Your own may differ! However, it's worth noting that using a tool with a strength control will give you the flexibility to choose just how much of the drummer's original swing you want left in there. This can be crucial in rock music, where putting something entirely to the grid is the quickest way to kill the groove. Remember, the drums become the basis of the timing and swing of the record! If they are too rigid, then every other element laid on top is likely going to feel that way as well.
Now, assuming we've timed the performance to our liking, we can move on to manual silence editing. This is an optional process. Some people like having bleed in their close drum tracks, which is absolutely fine, if they're after that aesthetic.
If, however, you want extra separation and cleanliness within your mix, you will want to do this.
A common set of tracks to start with here is the toms. Since the toms are generally a fill element, and generally only come in once every few bars, the ambient spill on them can be discarded. The idea is to chop the audio at the start of each transient, leave the entirety of the hit, all the resonance that you want, and chop again after that sustain has ended. You can create your own release envelopes here by using fade-outs of your desired length.
Some people will also manually silence hi-hat and ride tracks during this process. I personally don't, as I tend to not mind the bleed from hat and ride tracks. In a worst case scenario I will simply silence an irritating ride track during the mix, either by automating the channel mute on and off, or chopping and muting regions. The hat track is commonly so low in the mix that any bleed from it becomes largely inconsequential.
Now that this is done, we want to move on to placing some clean hits. Assuming that you were wise and sampled the entire drum kit at the very start of the session, you will have a collection of samples to work with, across all velocity layers.
The first order of the day is to get clean tom hits and place them at the end of every fill that has cymbals succeeding it. This way the cymbal bleed that frequently pops up when the drummer crashes on beat 1 of the following bar won't come through and mess up your balances. Use your own discretion here. Replace the hits where you see this bleed to be a problem.
After the toms are all done and consolidated you want to move on and repeat this process for weak snare and kick hits too.
When this is finally finished, and your drums are all consolidated, you have yourself a ready drum track with which to begin mixing! You will thank me for being so thorough during the preparation process, as this pays off in dividends during the mix, saving you both time and hassle.
Now we're at point A, so?
We're all ready to mix!
What you do here, and how you attempt to coax tone from the drums will depend in many ways on your personal inclination and the sound which the song demands. Rock and metal drum mixing do tend to differ in some fundamental ways, so I will try to outline common approaches to both whenever practical.
A common starting point for me is with the overhead tracks. The way these are treated tends to differ fundamentally between most rock and metal music, so I will split this part up right here.
Start by jogging a high pass up the overheads until the bulk of the drum hits and their body disappears. Your general intention here is to use the overheads as cymbal microphones. Excessive drum spill is undesirable (although not catastrophic). If you find you get a lot of snare poking through, you can use a simple brickwall limiter to take the peaks down. Otherwise you can set up a compressor that's keyed by the snare track, and this will duck the overheads each time it's hit.
Start by setting a reasonable high pass filter. For me this tends to be anywhere from 90 hz up to 170hz or so. A lot of this will depend on the quality of the room, which will determine the quality of low-end you'll be getting from the overheads.
Don't be afraid to crank some lows into them. Try to see if you can get some beef and body from the drums here. Combining this with the direct tracks will give you more of a 'pillowy' characteristic to the low-end.
Since we aren't high passing drastically here, we will likely be dealing with a fair amount of room mud. Check the area between 300 and 700hz to see what's going on. Usually a wide scoop around there is enough to clear things up and get it breathing.
Now that we have the low-end and mids out of the way, we can focus on some universal principles on processing overheads. Bear in mind that the techniques listed in the above approaches aren't necessarily exclusive to their respective genres. You can do a metal record and still keep a lot of drum sound in the overheads. Feel free to be adventurous, and define yourself.
The first thing to bear in mind is a potential trouble zone between 3 and 4kHz. This is part of the critical vocal range, and believe it or not, your cymbals will eat into it. A common approach is a wide scoop around this area as a starting point. Jog the cut around with your vocals going until you find that magic spot where they start to breathe through more naturally. In the mean time, make sure you haven't sucked all the life and presence out of your cymbals.
After this things become very circumstance-dependent. Most cymbals will have resonant frequencies in a few spots, and your job from here will be to tame these as much as possible without killing all the life in them. Similar principles apply here as with guitars and fizz. These are generally tight resonances, and being such they just love to eat into the clarity and front-to-back depth of your mix. You'll be amazed at how much clarity there is to be gained here once you refine the art of 'ringing out' the cymbals.
As always, be mindful of what does and doesn't need to be done. If you've tracked great cymbals with great mics in a great room, you may not even need to do anything more than the wide presence region scoop. So try to maintain perspective. Reference to your favourite records if need be for guidance.
In regards to compression, this is a personal matter and will depend from individual to individual. I personally do like to compress overheads slightly, and I do mean slightly. It's more for the envelope control than controlling the dynamics. If you find you have a very uneven overhead track, it may very well be a better idea to automate intensively instead of compressing, as the compression will bring along with itself negative effects, such as pumping, tearing cymbals, with increased resonances you may not want.
Last edited by Ermz : March 29th, 2012 at 01:50 PM.
|March 31st, 2011, 09:47 AM||#2 (permalink)|
Join Date: Apr 2002
Location: Melbourne, Australia
After the overheads are done, we have our perspective. Our 'whole kit', or our 'cymbal' sound as it may be. With these in place we can start throwing the other drums in one by one and making them work together.
The bass drum commonly has a particular EQ curve which is desirable for mixes. Many times this is achieved by scooping out the region between 200 to 400hz, widely, and heavily, followed by strategic low and high boosts.
This will vary from kick to kick, project to project, but in general it's a good idea to get a kick's thump centered around 70hz. This creates a good balance of low punch, whilst remaining high enough for most average stereo systems to be able to reproduce the bulk of it.
You might find under many circumstances that the kick has significant content between 100 and 200hz. This can be dangerous as it conflicts where your bass guitar is likely to lie. If necessary, feel free to take some energy away from this area, as it will immediately make your kick feel deeper, and your mix cleaner. As always, too much removed and you will thin the sound out too much, so be wary.
In regards to high-end boosts, the sweet spot tends to be between 6 and 10kHz for most kicks. It really depends on where you would like your high-end slap to lie, but a lot of the time the choice is self-evident as the kick will immediately sound better when you find it.
If after all this you find the attack of the kick is still buried you can try boosting within the presence region. Between 2 and 4kHz will lead to really present 'clack' type sounds, which in some ways characterise the infamous Slate kicks.
Ultimately, if you find yourself boosting a lot in the high-end regions, to the point where the kick is quite harsh and thin when solo, but drowned when in the mix, it may be advisable to look at the rest of your mix instead. Chances are that the guitars are too bright in the presence areas and are clouding the attack of your bass drum.
Fairly straightforward stuff going on here. You generally want a slow attack, moderately fast release compressor. Anywhere from 4:1 and 8:1 in ratio tends to work well on a majority of compressors.
The idea here is to add punch to the transient of the kick, rather than dynamic control. If the kicks are very inconsistent, your better options are to either automate them or sample blend/replace.*
The type of compression that tends to work well for adding punch to kicks is VCA-style. SSL channel strip compressors, Distressors, DBX160VUs - all great here.
The release time on the kick can at times work as an automation tool for you. Picture a record with a lot of double bass parts. If for some reason you don't want to manually automate each section of double bass parts in volume/high/low-end you can simply set the release of your kick compressor to overlap with each successive hit in the double bass run. This will have the effect of dropping kick volume during those runs, as well as somewhat dampening the transient.
A limiter on the back end can be good if you're dealing with a largely natural kick. This can stop some of the stronger hits overshooting with massive, slappy transients. I would set the limiter very conservatively, as it's almost too easy to kill the punch with one.
Kick Transient Design
In those circumstances where traditional compression fails and your kick is still lacking a little something, it might be a good idea to bust out the old SPL Transient Designer.
This is a great way to emphasize the transients on the kicks without the negative effects of over-compression.
Try it for yourself, and bump it to taste. I use it as a last resort, but at times it can give you just the punch you need.
The operative question to ask when augmenting a kick with samples is... how much? How much sample do you want, and how much real kick performance are you willing to trade for it?
For many metal genres the outright answer here can be 100%. At times it's desirable to simply have an extremely consistent, robotic kick. Bear in mind that this comes with an added need to automate closely in order to keep it gelling with the mix as the tempo and intensity of the performance varies.
As far as sampling goes, the kick is the most viable element to replace entirely and still fool the listener into thinking it sounds natural. This is because the kick is mostly about the feel and punch. Using samples gives more consistent punch, and as long as you keep a reign on the high-end from getting too robotic, you can get away with murder, so to speak.
Experiment with getting the sample to follow the dynamic curve of the original performance somewhat. Drummers will tend to emphasize certain downbeats and it can be quite a musical effect, if retained in moderation.
Ensure that your kick sample is in phase with the original kick (if blending). Sliding phase around can almost be like another EQ control, but the ideal spot with give you the most bottom end integrity. This is especially crucial if you blend a lot of samples into the one sound (we will get into frankensamples later).
This is a big one both for rock and metal genres. When mixing most rock music, you will find a need to lower the kick substantially for any slower, or cleaner verses, in order to have the headroom left to explode into choruses. For metal you will need to automate heavily in order to stop your heavily sample replaced kicks sounding overbearing during fast double-kick runs.
The principle here is quite easy. Mix the kick's standard tone during a section of the song that represents a state of 'normality' for the mix. For rock you may want to start at the chorus, which marks the point where everything is 'maxed', and downsize accordingly from there. With metal you may want to focus on parts not littered with too many intense double-kick runs, as they will inherently push you to subtract low-end and thin out the sound too much during single-hits.
For metal automation you will want to find spots where the kicks jump out and start to sound overbearing. You have one of what are essentially 3 options, and each may be applicable depending on how the kick is jumping out at you:
-If the high-end attack of the kick is becoming overbearing during fast runs, then simply automate a low-shelf EQ to subtract top-end every time this happens.
-If the low-end of the kick gets out of control during these sections simply automate a low-shelf EQ to subtract bass every time it happens, still leaving the top-end clarity.
-If the whole kick seems like it's just too in-your-face then simply automate the volume level down during these sections. Beware as it's almost too easy to lose the kick this way.
For rock automation your main tool will be volume. Simply move the fader around during the verses to find a spot where the kick sits well, and let it return to its punchy, normal tone during the dense, energetic parts of the song.
Some people like to use different samples for verses. Tamer sounding and more gentle, then kick into the punchy stuff on choruses. Others like to automate their compressors to ease off during lighter parts, and then put the squeeze back on during the dense parts. As always, the sky is the limit, and you need to experiment as much as you can, and find something that works for you!
The snare is likely the most intense, and difficult drum to get right. Not only does the tone originate from the drum itself, but is almost equally reliant on the space within which it's struck. The room tone is critical to good acoustic snare sounds. The snare is such a vital part of most beats and music, and in many ways is the most expressive of the drums, frequently containing the must subtle nuances in rolls, ghosts, flams and whatever else. This makes it notoriously difficult to keep sounding natural when augmenting with samples. The key word, after the base sound has been tweaked, is automation. But, we'll get to that in due time. Speaking of which, I hope you have a lot of it left because up to this point we've already covered over 6,000 words, and barely scratched the surface!
Most of us will be recording at least two snare tracks. The common standard is top and bottom. The top generally gives you the body of the drum, while the bottom is added underneath to emphasize crack and add some sizzle from the snare wires.
The main thing to note about raw snares – especially the close mic'ed tracks – is that they will sound like absolute crap. Even if you're recording the best snare ever forged on the fires of Hellhammer's burrito-ignited anus, it's going to take a bit of work to uncover that potential. So, this is normal. Don't expect to bring the fader up and hear a radio-rock ready snare with a second of room sustain behind it, and a crazy attack. Most snares will instead just be a muddy 'thwack', and the magic is something you have to help create.
You can choose to start with EQ or compression here – whichever sequence is more pleasing to you. I will begin by running through EQ first, but that by no means excludes you from dialing your compressor settings before starting to get surgical.
Top Snare EQ
Start with filtering. A HPF at 70hz is a good starting point. If you are going to take a lot of high-end from your snare bottom track, then you may want to lowpass here, anywhere down to about 12kHz. This may have the added benefit of rejecting some super high cymbal bleed.
From here, listen to the tone of the snare. If it's a lot ringier than you would like, then you've made a mistake in tracking and not used enough dampening on the snare head. In the case that you want a tight, dry sound from the snare, one of your main options is to go in with tight EQ bands and notch out the offending ring frequencies. With snares these are generally to be found anywhere from 300 to 900Hz. Essentially the entire low-midrange spectrum.
Once you're satisfied that the snare is ringing (or not) as much as you like, you get into the broad sculpting area. Snare inevitably need a fair chunk of low-mid reduction. A lot of the time they will be muddy across that same spectrum where those rings can be found. Try to isolate areas where the snare is particularly muddy, and take them out – but just enough to not kill the tone of the drum entirely. If the drum sounds 'tubby', then chances are your problem lies between 200 and 500Hz. If the snare sounds flat or two dimensional, then chances are the problem lies between 500 and 1kHz.
Now that we've largely 'de-junked' the snare, we're likely left with a more palatable, albeit somewhat weak and non-distinct sounding snare. This is the part where we start to add fidelity.
The bulk of the low-end from a snare tends to lie between 100 and 200hz, depending on its construction, the head used, and the tuning. You get two options here. Do you want to use a peak boost, or a low-shelf? Well, the answer depends on what sort of snare tone you're trying to achieve. A peak boost will generally give you a tighter, more controlled low thump, whereas a low-shelf will just lift the entire range of lows across the snare evenly. The shelf generally sounds 'fatter', but takes more space and is likely to excite some muddyness, whereas the peak boost will generally sound tighter, but smaller.
A good starting point for the low-shelf is around 100hz. Sweep it around and jog the gain until it sounds like it has enough thump for you. For the bell boost you can basically sweep it all throughout that 100 to 200Hz region, to find somewhere that hits home well for you.
So our snare finally has some body, and the mids no longer sound like someone deepthroating oysters. One thing we have left to do is to excite the highs somewhat, and make sure our snare cuts through our (hypothetically) dense arrangement.
Similar principles apply here, as with the kick. The 6 to 10kHz range is good for boosting the air. Just be mindful – this is where hi-hat bleed can kill you. One of the major drawbacks to processing raw snares is the amount of bleed that will start coming through. You may have to make a compromise here if you don't have as much isolation as you want from the cymbals.
The 2 to 4kHz range can be great for cutting through. If your snare is getting drowned, run a peak boost around here and see where it sits nicely.
Top Snare Compression
Once again, very similar principles apply as with the kick drum. Slow attack, with a moderately fast release, and a ratio ideally anywhere between 4:1 to 8:1. Attack times between 10 and 30ms will generally work well for you. Release times between 50ms and 100ms generally work equally well.
Be very careful with just how much you compress here. If your isolation isn't good you can end up with more hats in your snare top mic than actual snare. If bleed is a big problem and you need to add punch to the snare, try using a transient designer as your envelope shaping tool of choice instead of heavy compression.
Bottom Snare Compression
We're starting with compression on this one, because it's so vital to the overall shape and sound of this mic.
Bottom snare is a bit tricky, because the snare wires are quite prominent and get harsh easily. So what we primarily want from this mic is to add some sizzle and edge to the snare tone. We start with fast attack, fast release compression. You might want to hover around 8:1 ratio. We want to slam those wires down and increase their sustain substantially. I shoot for a significant amount of gain reduction here, until there isn't much transient left, and it's mostly a wash of sizzle, to complement the main (top snare) tone.
Bottom Snare EQ
Start by filtering, much like the snare top track. 90Hz is a good one for me here.
Feel free to dime the highs or high mids. The tone of the snares will determine what you can get away with. If you manage to make them sound airy, rather than harsh, then you can get away with blending more of them into the mix. As it stands however, think of this largely as an auxiliary mic to the top snare track. Blend in as much edge as you think it may need. A peak boost around 5kHz may get you started, but feel free to jog it around and see what works for you. You may find a shelf is less harsh and better for your style. Just watch out, as boosting too much broadband treble can make this get harsh, quickly.
Feel free to use this opportunity in order to boost more low-end into the snare. A peak boost anywhere between 100 to 200hz could be beneficial and add some extra body without exacerbating the muddyness found in the top snare track.
This is a fairly important one if using a dynamic source. Snare is usually quite dynamic, even with more consistent players, and certain hits can get quite harsh and 'pokey'. A quick way to get them under control is to use a limiter on the back end of your snare processing chain. Set it so it only acts on the offending hits, and leaves the majority untouched. It's very easy to castrate the punch of the snare here.
This is a technique which was popularized in recent years in order to get around the effect that brickwall chains in mastering have on snare transients. As the snare is traditionally the first thing to get lost after mastering, this was discovered as a shortcut way in order to retain its transients after maximization.
The process uses hard digital clipping to chop the top and bottom of the waveform in order for it to use less headroom. Now instead of a master limiter squashing the originally large snare transient downward whenever it hits, it instead lets the chopped peak through. The D/A converter panics because it cannot recreate the chopped waveform (a pure square wave doesn't exist in the analogue realm) so it essentially overshoots and creates a faux-peak which retains the illusion of a punchy snare on output.
Crafty? Yes! But very easy to over-do. Only use this sparingly – ideally while monitoring through a temporary master chain of your own. Too much clipping and you will castrate the attack of the snare.
Snare Transient Design
The transient designer can be a great tool for snares. It's a great way to emphasize the transient without the negative effects of compression (bringing up the cymbal bleed). You can even use the release portion to emphasize the life in the snare if it's exceptionally dead. Just beware that doing so can bring up unwanted floor noise and resonances you may dislike.
Ahh toms, my favourite! I quite have a thing for huge tom sounds. There is something really awesome about the way they sit, and the impact they add in a well-balanced mix. Personally, I avoid sampling these at all if possible. They can be notoriously hard to trigger accurately, and also have quite a pronounced fake sound if completely replaced.
Much like kicks, toms invariably have some mud in the lower-mids, and resonance in the mid-bass frequencies that can play havoc with your mix. How much, and where it's all centered will depend on many factors, ranging from the size of the tom, its construction, the head, the position and angle of the mic, as well as how it's hit.
A good starting point with toms is filtering, as usual. My high pass filters on toms tend to range between 50 to 80hz. This is a very subjective thing, and will depend on how you personally like to hear toms in the mix. If you want them shaking the floor, then you may want to consider lowering the filter down to 40Hz. In general I try to keep the higher pitched toms high-passed further up than the low-pitched ones, in order to let each resonate somewhat in its dominant zone.
Let's look at the low-midrange now. There is generally a collection of mud between 200 to 400hz on toms. A wide scoop normally does wonders down here in clearing up their tone, and how nicely they play within your mix.
The next area to look at is between 600 and 800hz .There is usually a flat or '2 dimensional' quality to toms stemming from here – especially if the room was small. Feel free to take away some content here in order to let the toms breathe more. Sometimes you have to get fairly brutal with your subtractions in order for them to sit right. Toms can have very hefty build-ups of undesirable mid frequencies.
After this you will find that 6 to 8kHz is a good range to start with for high-end boosts. It's essential that you have the overheads in while doing this because they provide a lot of the air to the toms, and you may end up with something brighter than you intend by only soloing.
If the toms still don't cut, then consider our familiar 2 to 4kHz range to get them through. This will boost stick attack and really help get them through that wall of guitars.
Lastly, if you find your toms are still overly resonant and not sitting how you'd like, take a look at the 100 to 200Hz range. They will frequently have resonant peaks here which can be helpful to subdue. Using this in concert with a low-boost will help create a larger, and less obtrusive tom sound.
For low-end, either try the default shelf at 100hz, or simply run a peak boost between 50 and 120hz (depending on the size of the tom, and where it needs to hit).
Nothing out of the ordinary here. Similar principles to kick and snare. Slow attack, moderate-fast release, around about a 4:1 ratio. Good compression range for toms tends to be between 3 and 6dB.
Limiting toms is not something I personally do, as I like them to come through with full force. However, I frequently set up a secondary compressor on the tom bus to help control the transients a little. It's normally an FET-style compressor with fairly fast attacks, so that it both adds density and controls errant peaks to a degree.
For your average metal mix this is going to be one of the most non-essential tracks you have. It can be used to add a sense of glue and 'togetherness' to the kit, but otherwise will only serve to add mud and junk.
For a rock mix, this microphone is almost essential. It contains the crux of glue for your kit, and will also be a potential source of extra low-end for your drums.
The very biggest problem you will run into with most room tracks is cymbal spill. They will seem like they're coming from everywhere. As such, you may want to try a substantial low-pass on them.
Think of this as an auxiliary set of tracks. You are trying to create a background wash, and a sense of glue to unite the kit and give it a greater sense of density. Most of your work with EQ is to remove the mud that builds up in the room, take out the cymbal crazyness, and emphasize the lows.
Start with wide cuts throughout the low-mid spectrum in order to clean the tracks up. 300Hz is one I find myself going back to frequently to get junk out. Another can be up around 700Hz, to kill flatness.
The important part here is to crank the lows. Turn up a low-shelf around 100hz and let the room mics fill out your drums' low-end. The kick will take on more of a dense, pillowy characteristic, and this will help you in filling out sparse drum parts with extra tone and sustain.
Great starting point here is the 1176, or 1178, all-buttons-in mode. Attack anywhere from noon to 9 o'clock, release full clockwise. Slam the room until it's vibing just right for you. Be wary however, this is where the cymbals start to become a real problem. After you're done with your slamming, you may find you need to get more extreme with your low-pass filter, or simply adjust to the idea of using the room mics for some of your cymbal tone.
Cymbal Spot Mics
The general idea with these is to compliment the overheads. You can leave a bit more body in the direct tracks and bring them up under the overheads to create a sense of solidity. With spaced pair overheads these spot mics can be fantastic in defining a solid spot in the stereo image for a cymbal to sit.
Always try to process these with the overheads going, as they are solely support mics. High-passes on these can range from 3 to 500Hz. After this you might find yourself removing some mid junk from 5 to 700Hz. Always be mindful of the critical 2 to 4kHz vocal range, and make sure these aren't eating into it. From here it really depends on what the overheads need. If they are excessively bright, then it may be a good idea to keep these direct tracks a bit duller. If you need a bit more sizzle, then perhaps consider pumping the highs.
Bear in mind that spot mics tend to collect a lot of junk. So don't feel bad if you need to cut substantial mids on these tracks, or have to do a few, or even several notches to take out high frequency resonances.
Sometimes you're well off using a low-pass if simply trying to add weight to the overheads. Taking out everything above 12kHz should take care of any sizzling which can build up, and leave you with the meat of the cymbal sound.
Compression on these is entirely to taste. I personally tend to enjoy slower attack compression to add some transient and attack to the cymbals. This can be counteracted with a limiter on the back end to stop cymbals like rides getting excessively harsh on certain hits.
General Drum Mixing
Reading through the section above, you may get the idea that processing drums is a very linear, disconnected affair – but nothing could be more from the truth. Drums are extremely interactive, and changes made to one track will many times affect the perception of the others. While mixing, I am constantly thinking about half a dozen of these interactions at any given point in time, and you need to develop your ability to do the same. You need to hear the problems, isolate where they are, and fix them, thinking in multiple dimensions.
Parallel Compression – Yes or No?
Parallel compression can be very handy in adding sustain to drums. It typically doesn't work as great on cymbals as it does on shells.
Snares are notorious for their love of parallel compression. It can be used to add a great deal of sustain to their tone, and immediately adds a 'rock' character to the sound.
Kicks can benefit from parallel compression too, as it rounds off and enhances the thump on the low-end.
It should be needless to say that if you're mixing very fast, technical metal music, you'll frequently be better off circumventing parallel compression entirely. It thrives on increasing the density on drums, whereas with these technical and/or fast genres of metal you're after tightness and precision. In this case the compression would be a quick way to turn your drums into a complete mess.
It may seem like we haven't spent much time talking about the role of samples in this guide. Up until this point, that's true. I wanted to largely focus on how to get the most from the raw drum sounds before we proceeded to augment them (if necessary/desired) with samples.
One of the most exciting aspects of drum mixing is preparing custom samples for each project. Even if you have a bank of 10 samples, their potential combinations with one another open up a whole array of possibilities for you.
What I would advocate doing is preparing the samples for your project ahead of time. For instance, I commonly run a 'premix' session before I begin the mix on a record, with the entire album loaded up in there. I will use this opportunity to get a rough mix going with the drums, and then experiment with which samples may work in tandem. Once I find a great sample that forms the base of the interaction with the raw drum, I will then go on and try to keep adding to it by finding other samples that add something desirable to the tone. So it's quite common for instance to mix a very low 'thud' snare for mega low end, with a higher piccolo snare for cut. Both contain elements neither would be able to possess all by itself in the real world, so what we're trying to do here is create a 'frankensample' which has the best aspects of everything for our record.
A common question is 'do you just line all the samples up and process them together, or do you filter each individually for only its desired characteristics'. My answer is: whichever you want it to be. If you are set on filtering only what you want from each drum, that's absolutely fine, and will give you a lot of precision in the sound you tailor. Personally, I tend to stack them, full-range, until I notice something about an individual sample that bothers me. Let's say I like the attack of one kick, but the overly resonant low-end isn't gelling with the rest of the drums. Well at that point you simply high-pass until you're left with only what you want from that drum.
After all the samples are lined up and worked out, if I have the time I will endeavor to process each individually. Essentially de-junk it and focus it into its dominant zone, and create a clean, mix-friendly sample. The other option, of course, is to just let them all combine, and process them on a collective snare bus. Ideally, I do a bit of both, for maximum flexibility.
Samples really let your imagination guide you, and create a whole new layer of fun to drum mixing. This is where things can start to sound really huge.
Just remember, if you don't like the way your sample blend is sounding, don't reach for the EQ! Either re-balance them in level, or mix up some different samples. It's always better to mix with the raws and levels, rather than processing. This way you can use EQ and compression only as the final element in getting the tone where it needs to go.
This is one that comes up quite often. Sampled rooms are a true wonder and a blessing for those of us working in the digital age.
As we found before, one of the biggest limitations to our acoustic room sounds is the cymbal bleed that creeps up and messes up our mix. With sampled rooms, we no longer have that problem. We can process them to our heart's content, and extract only the ambient qualities we need for our shells.
Another great aspect to using sampled rooms is mixing and matching. The room ambiance doesn't have to be of the same snare you have going in your mix. You can combine some really thuddy, low, direct snares, with a very high room wash. It's a mix and match exercise until you get to where you want to be.
The processing principles are similar to acoustic rooms, except you no longer need the low-pass. You can compress all you want in order to get all the body and ambiance into your direct drums that you require.
When the time has come again, and I'm doing one of 'those' mixes (the completely inappropriate variety with rock processing on a metal record) these sampled rooms become a huge part of my overall drum sound, which are enhanced even further by parallel compression on the drum bus. It turns the drums into a living, dynamic being, that you're attempting to reign in with everything you have. A world apart from the 'tick tock' modern metal drum sound that's become commonplace.
Last edited by Ermz : March 31st, 2011 at 10:03 AM.
|March 31st, 2011, 09:48 AM||#3 (permalink)|
Join Date: Apr 2002
Location: Melbourne, Australia
This one is really all down to the individual. In general drums tend to like colourful plate reverb. Plate reverb is commonly really dense and vibely – audible as part of the music itself. It can really help snares come to life, above all.
Beyond this drums can take anything from room to hall reverb. If your tracking room was less than ideal, you can add a sprinkle of reverb to the overheads and room microphones. After this you can beef up the close mic tracks.
Dense mixes commonly eat reverb up, so you can have a fairly wet drum mix and it may not translate as such in a dense rock or metal arrangement. With that said, in mixes where I want a lot of ambiance, I tend to favor rooms and room samples over artificial reverb. The artificial reverb only tends to come in for the long tail that isn't possible in the real world.
You can colour and trigger verbs by sending different samples to them. One trick is to use a snare sample solely for the purpose of triggering the reverb a certain way, but otherwise having the sample itself muted. Beyond this you can feel free to EQ the reverb return to your heart's content. It's another mix element, like anything else, and can be shaped to fix the scope of your landscape better.
As a footnote, make sure that you always have a high pass on your drum reverbs. This will help keep your drums clean and intact. Anywhere from 150 to 200hz should do.
This is a summary of many of the frequency regions we spoke of, and how they relate to most drums. As always, no hard and fast rules, but drums do generally have common problem zones that can be tackled in a similar, and uniform fashion each time.
22 to 70hz: The lowest of sub frequencies. Can be filtered out on snares and toms (though toms can go lower). Kick punch will ideally lie somewhere here. Most commonly between 50 and 70hz.
100hz to 200hz: The location of snare bottom 'thump'. Can be boosted for more balls on snare. Can contain mud in kick and tom tracks.
200 to 400hz: The 'mud' region. This is almost universally true for all drums, from kicks all the way through to the room tracks. Scooping in this area will aid in clarity.
600 to 800hz: The 'flatness' region. Scoop this in order to add more depth to drums.
2kHz+: Everything from here on boosts stick attack or cymbal presence.
2kHz to 4kHz: Boost on kick, snare & toms in order to cut through dense mixes. Cut on overheads to stop conflict with vocals.
6kHz to 10kHz: Good for adding top-end clarity to anything.
10kHz+: Sizzle and air. Watch out, things can get somewhat harsh and weird up here.
And that essentially concludes the drum guide for the time being. It was a mammoth task to write all in one evening, and doubtless there are some issues here or there that I will amend in the coming weeks. Thanks again for bearing with me for the second instalment of this guide, and I hope you managed to pick something up.
Last edited by Ermz : March 31st, 2011 at 10:04 AM.
|March 31st, 2011, 11:27 AM||#11 (permalink)|
Join Date: Sep 2008
Location: London, UK
I appreciate the tip about sampling toms and using them to cap off fills in order to hide the upcoming bars cymbal spillage. Never really done that before, but definitely going to try it!
One thing I'd say is that I find a bit of harmonic distortion can help on giving toms a bit of presence, but only a little!!!
|March 31st, 2011, 11:58 AM||#12 (permalink)|
I <3 Violent Euphony
Join Date: Oct 2008
God DAMN. These 'guides' you have been writing are fucking AMAZING dude. Props. I'm SO MUCH MORE of a knob tweaker in general. twist til it sounds good =D
definitely printing this out and saving them all for future reference!
Destroy rather than create what's meaningless
CHARLES J'S - "GRISLY GALLOWS KIT: and "DEATH DRUMS BUNDLE" - AVAILABLE HERE!!:http://charlesj.bigcartel.com/
CONTACT CHARLES J OR PHIL PLUSKOTA @
SONIC ASSAULT STUDIOS
|March 31st, 2011, 11:58 AM||#13 (permalink)|
Join Date: Jun 2010
Ahh, the part I've been waiting for. Lots of good information here, as with all your write-ups.
There's quite a few things that I do differently, but I suck at explaining things.
One big one, though, is (with a dull snare that you can't seem to get any attack from) to side compress the shit out of both snare mics. I mean, like 8:1-10:1. I usually use an attack around 20-50ms (longer = more crack) and a release around 100ms. That will give the snare a considerable snap/crack.
As always with anything in this field, your mileage may vary considerably.
|March 31st, 2011, 12:49 PM||#15 (permalink)|
Join Date: May 2006
Read some of it, we agree on many things (I just couldnt read everything...12+h daily in front of the computer mixing + share of personal problems to deal aint doing good to my head). Thanks/congrats for it!
One of the things I struggle the most is tom sound. Drummers always hit it like a little lady, so it ends up sounding like crap - and I hate sampled toms. Anyway..
Mixing, editing and drum programming - PM for quotes
Last edited by jangoux : March 31st, 2011 at 12:51 PM.
|March 31st, 2011, 02:21 PM||#17 (permalink)|
Join Date: Sep 2008
Location: London, UK
|March 31st, 2011, 03:56 PM||#20 (permalink)|
Join Date: Jun 2010
Location: Belfast, N.Ireland
Just finished wow, what a long read, great though definatly picked up a few things im guna try on the few live drum projects i have, as with the last its great to to have your perspective of things and ill definatly be reading on and keeping my own documents filled with these tutorials for future referance, this one clocked in at 8,500 words, longer than anything ive had to write for college in the last two years really appreciated
Start Together Studio
|March 31st, 2011, 04:25 PM||#21 (permalink)|
Join Date: Nov 2008
Location: Munich, Germany
Thanks a lot - again. Made me realize I have to put way more attention to mixing drums.. Using programmed drums with Slate samples made me really lazy.
Really looking forward to the upcoming articles.
|March 31st, 2011, 04:34 PM||#22 (permalink)|
A Venezuelan in Europe
Join Date: Jul 2008
Location: A Coruña, Spain
Ok I'm printing this out, my eyes will bleed if I try to read that on the monitor.
Inmense respect to you Ermin, to actually take the time to do this, we love you!
Weekly Favorites - Currently Listening
|March 31st, 2011, 05:22 PM||#25 (permalink)|
Join Date: May 2010
Location: The Netherlands
I really enjoy your Systematic Mixing Series. Is there any way I can donate, to stimulate your great effort? I can't make you a millionaire, but hey, even the little bits count!
And, do you mind if I save these tutorials on my computer in a document. Ofcourse, I won't distrubute it or whatsoever.