Engineer Richard King has brought the art of ensemble recording to new heights in both classical and folk/pop spheres.
Since their inception in 1959, the Grammys have included two Best Engineering categories: Classical and Non–classical. During the ’60s, Capitol Records engineer Hugh Davies became the first person to win both awards, a feat that was to remain unmatched for more than 40 years. The engineer who finally repeated this achievement in 2012 was Richard King, a man with more than 12 Grammys to his name, and who was nominated in both Best Engineering categories again this year for his work on the St Louis Symphony Orchestra’s John Adams: City Noir and Chris Thile & Edgar Meyer’s Bass & Mandolin. Following a 15–year stint at Sony Music, he now runs his own independent business RK Recording (www.rkrecording.com), and has worked with top orchestras and classical stars such as Emmanuel Ax, Joshua Bell, Renée Fleming, Hilary Hahn, Yo–Yo Ma and Itzhak Perlman.
Despite this impressive track record, however, he’s rarely been interviewed about his methods. So when I discovered that he’d recorded, mixed and mastered Yo–Yo Ma’s The Goat Rodeo Sessions — quite simply one of the most glorious–sounding acoustic records I’ve ever heard — I made it my mission to track him down, and he graciously agreed to share some insights about the recording process.
Like many other engineers working with acoustic music, King typically relies upon a main stereo (or surround) microphone array to capture an ensemble’s overall sound, supplementing this with spot mics for individual instruments or sections. For his main array, omni mics are firm favourites. “If I’m in a room that’s pleasing and will work with me as a major component of the sound,” he explains, “then I want to use an omni system so I can include that as part of the overall timbre and sound and space. For me, it’s a quick and easy tool for getting what’s happening in the room to translate over the loudspeakers in the control room.
“But you have to take the good with the bad. You have to be comfortable including all the breathing and the clicking and the noises of the instruments. You’d go crazy trying to remove all of it afterwards. The whole concept of recording with omni microphones is that you just have to embrace all of it as one sound source, and get over the noises. But I’ve never really been that disturbed with noise in performance. Key clicks on a bassoon, for instance — it’s just part of what’s happening. It’s the same with the mechanics of a piano. As long as it’s not too distracting, it’s just part of the instrument. You can try to minimise it, but I never try to remove completely all of those artifacts.
“You have to embrace all the leakage as well, but you still have to end up with some control. To really pull this off, ideally the musicians shouldn’t have headphones on. They should listen acoustically, so they’re hearing the natural balance you’re recording, and can adjust to it in real time. If everyone has their own headphone mix, then everybody has a completely different concept of the balance and then you can’t rely on them balancing themselves; at that point it all falls back on the engineer, and you need much more control over the leakage. I usually keep the musicians close together, again so they can balance themselves without wearing headphones. As soon as you spread them out too much then they want to have headphones on for timing and clarity.
“The musicians should also come into the control room and listen. You have to do a lot of playbacks to make sure they’re aware of what the balance is over the microphones. Sometimes after listening, and with very little discussion, they’ll go out and do another take and they’ll naturally adjust to what they’ve just heard in the control room. I always have spot mics to help refine the balance, but it is still mostly locked in on the main mics if they’re omnis. You don’t have many options for adjusting under this approach, so the artists really have to be on board.
“Another thing about working with omnis is that the musicians have to be open to adjusting their positions in the studio. Quite often they’ll come and listen and say something like ‘There’s too much viola!’ but rather than have the violist concentrate on playing a little bit softer all the time ,you can just physically move them back away from the main mics a little bit and improve the balance in that way. They have to be OK with shifting here and there. And of course we mark all the chair positions, so every time the musicians come back from a break I make sure everyone’s back on their marks.”
The exact configuration of the main omni mics will frequently depend on the size of the ensemble. “Quite often I do just a left–right pair of omnis,” says King. “Sometimes I’ll do a left–centre–right. It depends. If it’s a small group and I know I’m going to come in pretty close, I can just get it more even by having a centre mic as well as the left–right. It’s better for coverage. Goat Rodeo, for example, was a left–centre–right ‘tree’ — three omnis pretty low, about six foot high — and the group were roughly sitting or standing like a string quartet would sit. It wasn’t a real Decca Tree, though. The centre mic was quite a bit deeper, closer to the bass. I keep that centre mic quite a bit lower in the mix, especially if I place it more deeply, so I’m down about 10dB from the left and right mics.
“With larger groups I’ll focus on two mics, usually about four or five feet apart, and with orchestra I like to put up a pair of omni outriggers for coverage as well. Everything’s panned hard left–right, so if I want more width and blend I just add more gain from the outriggers. Outriggers for me are tighter and quite a bit lower than the main mics. What I really want is that full string sound, and to favour the front half of the orchestra, because in the mains, quite often, the strings will be a little bit under and the brass will be strong. If I position the outriggers close to the strings I get a lot of core string sound there that can help to tip the scales back in the right direction.”
To support the main pickup, Richard King typically uses plenty of spot mics. “With orchestra, for instance, instead of a stereo woodwind pair, I’ll do each section of the woodwinds on a separate mic, and quite often bass clarinet and contrabassoon and English horn separate. So I’ll have seven mics for woodwinds usually. But then I sneak them in at a lower level than I would if I just had two mics in stereo. I spot all the strings — I try not to use those mics, but the bass spot I’d use for sure. For me, the best string sound is on the omni outriggers. With cardioid string spots, unless you have a lot of mics on each section, it’s just a weak, thin sound.”
Similarly, the core Goat Rodeo ensemble was also comprehensively spot–miked, the overall sound and balance decided in collaboration with producer Steven Epstein. Again, though, these signals were mixed with restraint: “I’d say we sneaked in the spot mics sort of halfway,” remarks King, “so the main sound was still that left–centre–right tree.” Most of the mics in this case were, as you might expect, directional, including a Neumann KM140 cardioid pencil mic and a Royer R121 ribbon on Stuart Duncan’s fiddle; another KM140 and a Chameleon Labs TS1 valve mic on Chris Thile’s mandolin; and another R121 with a Neumann M147 valve mic for Edgar Meyer’s bass. More surprising, however, was that Yo–Yo Ma’s cello was captured with a pair of DPA 4006 omnis. “On Yo–Yo, the best sound for his cello is omnis,” says King, “even as spot mics. I come in quite tightly, maybe two feet out following the angle of the bridge, using a pair of 4006 capsules on a stereo bar about five inches wide. I leave the pans full left–right, so I’m never adding right–channel signal to the left, which means I don’t have to worry about any phase cancellation with the two omni spot mics being so close together.
“The reason I use a pair for the cello is because I’m going to have a lot more leakage on these omnis than on any of the other spot mics. If I need to add a lot of cello support — let’s say he did something very softly, but it’s a solo and I really have to raise it up — having an omni mono source for him means all the leakage from the bass and the fiddle on each side is going to come up in the same place as the cello, and it’s going to distort my main image. With the two omnis close together, the cello image is quite mono, but the bass leakage and the violin leakage either side stays stereo. Yo–Yo’s sitting leftish [in the main pair’s image], at around the 10:30 position, so I’d have matched that by just lowering the right spot mic’s fader slightly.
“If you keep the leakage in stereo you can get away with a lot more. I discovered this doing concertos: if the orchestra’s blaring away and you want to crank up the solo violin, you don’t want this big mono clump of orchestra leakage in the centre. I’d use spaced cardioids for the violin, but again on a stereo bar, about five or six inches wide, and both facing the instrument in parallel. Another reason I do that for live concerts is that if you try an X-Y and you lose one microphone in the middle of the concert, you’re left with a mic that’s off axis to the instrument, but you need both mics to get the right sound. Having two mics parallel to the instrument, on the other hand, if I lose one, I still have one mic that’s on axis.”
Refining each instrument’s recorded timbre is all about what goes on in the live room, according to King. “I’m the kind of person who’s moving the microphones around to get the sound I want, rather than going straight to the console or plug–ins and starting to change the sound there. You sometimes see situations — and I don’t know whether it’s laziness or lack of training — where the microphone gets placed in front of the instrument and that’s the last time the engineer’s in the room, and the rest of the time they’re fighting the sound at the console. Whereas I’m always jumping out of the chair and running back and forth from studio to control room. If you’re recording with omnis, they have to be in the right place, otherwise it doesn’t work, so you really have to spend time working on that. For me, EQ should really be the last resort. You always end up there, for the super–fine adjustments, but I try to do the large brush–strokes out in the studio, moving the mics, maybe altering the environment. Those things get you in the ballpark more quickly.
“I tend to set the positions of my spot mics based on the sound I’m getting from the individual instruments, and then assess the interaction between the mains and the close mics in terms of time and phase. I certainly check polarity on low–frequency instruments, between the bass mic and the mains, but because the other spot mics are usually in the mix at a lower level than the main microphones, the comb–filtering and all of those artifacts are less of an issue. You have to manage it, of course, and if you start fading something up for a solo, you have to be aware that you might start introducing comb–filtering. Worst–case scenario, in the mix process I will delay a close mic to line it up with the mains if there’s a phase issue I can’t solve with the polarity switch. At the end of every recording session I’ll measure the actual distances between the mics, so I’ll have a starting point, but I tune the delay times by ear. With singers it’s sometimes a concern, because phasing is so much more apparent on voice than with instruments.
“Sometimes, funnily enough, you solo up your close mics and they sound terrible, but you listen to them integrated with the main picture and they work very well together. I’m always listening for the right character on the close mics, but quite often they don’t sound like much on their own; they just have the right material that’s missing from the main mics, to introduce into the mix. Quite often people ask ‘What mic did you use on the cello?’, but it’s not just about the cello mic. The cello sound was mostly provided by the main mics.”
Despite his preference for omni mics, Richard King takes a pragmatic view to working in unfamiliar acoustics: “I try not to presume too much. If I walk into a room, I try not to think about what it might sound like. I’ll wait until I hear it over the microphones. You can spend so much time working on something based on a hypothesis, and then be completely wrong. It’s just a waste of time. If I go into a hall or a room where they have variable acoustics, I always say to open it up all the way and give me the most ‘live’ you have, and I’ll tame it down from there. Especially with omni mics, where I’m trying to get a natural sound in a space, I want to hear as much of the space as possible at the beginning. And then if it’s too much I come in closer with the mics or start messing with the room. But I don’t waste time trying out something that I can tell will be a microscopic change. Let’s say we’re in a huge room and it’s too reverberant and someone says ‘Well, there’s eight chairs in the hallway, we could bring those in and put them along the front.’ You’ll soak up 0.5 percent of the reverb in the room that way. I can get rid of 20 percent of the reverb by moving my microphones a foot closer to the instruments!
“I’m not so crazy about using gobos. If you put a gobo between the microphone and another player, you’re going to hear a little bit less direct sound from that player, but you actually end up getting more reflected sound from that player, so in fact the leakage is worse. When you think about it, if you have direct sound leaking in it’s more clean than something that’s been bouncing round the room a half a second. For instance, if you’re putting drums in a booth, you’d better make sure the booth’s really closed up, because if you build a booth and it’s a bit makeshift, you’re going to get bleed coming out of that, and that could almost be worse than having them just in the room close to your main pickup.”
Where gobos can be very useful, though, is in improving the listening experience for the performers in the studio. “On Goat Rodeo, for instance, any time the mandolin was doing anything rhythmic you could really hear the walls as part of his sound. And the pick sound Chris Thile was getting off the plectrum, he could hear it coming back from the wall and that was driving him nuts. So there, yeah, I brought in a gobo about five feet tall and 10 feet long, I rolled it in between him and the wall — closer to the wall — and I hung heavy packing blankets over it. And sure enough he’s playing and says ‘Wow, the wall’s gone! This is great!’ It was more for him in the room than for what we were hearing in the microphones. He was much more comfortable, and we got what we wanted as well.”
In cases where the room turns out to be inappropriate in some way, King might switch the main array to cardioids to reject it, although merely repositioning the omnis can also prove an effective solution. “For instance, on the Silk Road Journeys album I did with Yo–Yo Ma,” he recalls, “we were at the Hit Factory in New York, which is a big room, but it’s kind of a pop room. The walls come back very quickly. It’s a short decay, but the room is very live, and it wasn’t really working well with the style of music. It was a little bit bright, I guess, with too much high–frequency reflection. So I stuck with the omnis, I just came in very, very low, and very close to the group, and was still able to get the ‘glue’ that you’d expect from an omni pickup. Super–wide frequency range, lots of friendly low frequencies. And with a lot of those instruments it was nice to have an intimate sound anyway and get good full bottom end in a natural way, so it still worked well in that case, even though I was in quite tight with the mics.”
But he also has another trick up his sleeve for dealing with less–than–ideal acoustics. “With omnis you’re collecting all these reflections, and if their quality is unwanted you can mask some of the ugliness with the right reverb,” reveals King. “In this case, I run a reverb with no pre–delay, because I want the reverb to start immediately with the direct sound, so the reverb is active before those reflections have come back from the walls. So let’s say the reflections come at around 10–20 ms or so. My pre–delay is usually zero, but certainly lower than 20ms. So you hear the direct sound, my artificial reverb, and then the room, in that order. It just takes the ear away from the room reflections. That works really well. But if the room were beautiful, then I’d do the opposite. I would push the pre–delay on my reverb to come in after the room reflections, so that the ear would be more aware of the natural space.”
Given how much of the sound of Richard King’s productions comes straight out of the mics, mixing is mostly a question of refining the levels of the spot mics in the balance; other processing will typically be minimal. “On Goat Rodeo, for instance, we took a little bit of low–mids out of the cello and just bumped the low end on the bass close mics 2–3dB, and really that was it for EQ. A tiny bit of compression on the bass close mic. A little bit of bus compression, 1.5:1 ratio, just tickling a little bit.” Automation naturally plays an important role in this, though, highlighting solos and performance details to best showcase the music.
There’s one mixdown effect he’s not shy with, though. “I have a reputation for using a lot of reverb,” he says, “and I do like reverb. It’s a lot of my ‘glue’ and the cushion in the sound, and I’ll usually err on the close side with my mics while recording. But I’m very picky about the quality of the reverb. If it’s the right quality, then I’m happy to have a lot of reverb going. A lot of that is designing the tail of the reverb, the way it decays, being sure that it’s not getting in the way. I want a reverb that’s active and busy around my direct sound, but when the direct sound goes away on an ending chord or a pause in the music I don’t want the reverb to be doing somersaults! I like it to gracefully drop out of the way in a more subtle way than a stadium reverb or car park.
“The first thing I look at is the size and length of the reverb. I work a lot with Lexicon, but most algorithmic reverbs have the same functionality. For the slower tempos, I’ll add a little bit more return, and I might just dry up a quicker tune by having less, but I wouldn’t so much change the length. Then I go to the timbre of it. I don’t like to use a separate EQ on my reverb return, I prefer to use the timbre controls within the reverb, so I’m looking at the high–frequency cutoff, the roll–offs within the reverb.
“From that point, the last and most important thing is to look at the shape of the decay — Lexicon actually call it Shape, but other manufacturers have other names for it. With a higher Shape number, when a pause comes in the music, the reverb actually hangs a little bit before it starts to decay. With a lower Shape number the decay’s more linear, and when the music stops the reverb just gets out of the way quickly. That’s the way I design my reverbs, so I can use a lot, but then you’re not aware how much reverb’s actually there when the music stops.
“If it’s a record like Goat Rodeo where everyone’s in the same room, I tend to have one reverb. If there’s a singer, quite often I’ll have a separate reverb for the singer only. But I might actually have all the same parameters set up for the vocal as I do for the rest of the group, but I’ll just run them into separate machines. I will be running reverb at the recording session as well, even if it’s not exactly the reverb I end up using at mixdown, it’s something close so I have an idea of the final product. I never just place mics and say ‘Oh, we’ll just listen to it when we get home!’ We’re hearing roughly the final mix when we’re listening at the session, so we know we have things in the right place.”
If you’re remotely interested in recording acoustic music, you’re probably already thinking of experimenting with some of these methods yourself. However, most SOS readers simply don’t inhabit that same stratosphere of top–tier musicians, venues and equipment, so I ask King if he has any specific tips for those on a budget. “I think omnis are an incredibly powerful tool for someone who can’t afford to buy a lot of microphones,” he says. “If you have a decent room, you can get an entire ensemble sound on two or three mics, instead of using six microphones in cardioid, so there’s good value in learning this style of recording, that’s for sure. The more you isolate, and the more microphones you add, the longer it takes to mix and get things to recombine after the fact. Managing and using the ‘glue’ and leakage at the session just makes it faster to get to the end product.
“An urban legend that should be undone is that a room should be covered with blankets and made as dead as possible, to remove all the reflective sound. What happens is that you remove the high–frequency reflections but you still have tons of low mid–range, and in fact, it sounds worse than if you’d just left it alone. I would encourage new engineers to just go into their living room and put the instrument somewhere in the room where it seems to be well–balanced, and put a microphone on it and listen. Start from there. And then if they discover they’re hearing a lot of sound off one wall in particular, OK, maybe put something on that one wall, but don’t go crazy with treatment that isn’t really broadband.”
Some corners, however, can’t be cut. “Inexpensive microphones always sound inexpensive,” King insists. “What I tell my students is that you can get away with a lot on your guitar amps and you can use a decent direct box for the bass, but when it’s time to record your vocals, go out and rent a decent vocal mic for a day if you can’t afford to buy one. I think the most important things are the microphones and the loudspeakers — the two ends of the chain.”
For even more in–depth information about Richard King’s recording techniques, check out his forthcoming Focal Press book, Recording Orchestra & Other Classical Music Ensembles (ISBN 978–1–138–85454–3).
For surround recording, Richard King will often supplement a forward–facing left–centre–right constellation of DPA 4006 omnis with a further pair of 4006s for the surround channels, the latter fitted with DPA’s acoustic equaliser spheres. The rear mics will be a few metres further away from the ensemble, facing back into the hall. “The ball makes the microphone more cardioid at higher frequencies, above around 1kHz,” explains King. “So at low frequencies you get this great correlation, except for the time delay between the rear mics and the main mics, so you get a lot of envelopment, but it’s a very stable image, because your source stays in the front speakers. And then the high–frequencies are decorrelated, so again you get the sense of the source sound really staying in the front speakers.
“It works really well. The first person I heard of using it was Erdo Groot, he was at Philips at the time, now he’s at Polyhymnia. He played me some of his recordings where he was using the acoustic equalisers for the rear, and it worked so well that I just adopted it. But I don’t like them for anything on axis. I don’t use them on the main mics, because it distorts the natural timbre of the microphone. It’s too bright on axis, a pretty big boost in the presence range, and it’s not pretty, it’s not a smooth boost. However, if you take the Schoeps MK2H and put their ball on it — it’s smaller — that actually sounds pretty good on axis, and it’s a better emulation of the old Neumann M50. The original idea with the DPA was to put the ball on and get the M50 characteristic, but that just didn’t work, whereas the Schoeps is much closer, and smoother, and for film the Schoeps with the ball is a much better replacement for an M50 than anything else, for a Decca Tree certainly.”
Richard King: “I’ve had more success using DPA omnis than any other omnis out there. That’s not to say I don’t like other mics: the Schoeps omnis are fantastic, for instance. Sennheiser has a new one, the 8020, which is very good and comes very close to the DPA for me. But there’s something about the colour... You know, every microphone manufacturer says their mics are perfectly flat and transparent, but every mic has a colour. For me it’s subjective, and the colour I like the most is what I get from the DPAs. They’re pretty flat. There’s a bump up around 18kHz or something — where, these days, I don’t hear that much of a difference! There’s a colour there, but it’s a positive addition to the sound, I think. They’re super natural–sounding, especially in terms of the space and the depth. Of all the omnis out there, I think they sound the most omnidirectional, in other words, if you put up a DPA and a Schoeps at the same distance from a source, both omnis, you’d hear more room from the back of the DPA than you would in the Schoeps.
“I started my career using the high–voltage DPAs, because we had them at Sony. They’re very good mics, obviously, except that you need to run with an external power supply, and then it gets tricky if you’re hanging your mics and that sort of thing, so I’d say in the last 10 years I’ve really been focused on the 4006. And it’s the transformerless one I prefer — it’s a little more open, and it’s a little more forgiving when you come in close on high strings, whereas the old ones used to get a little grainy.”
During recent years, Richard King has been scaling back his production work to devote more time and energy to research and education in his capacity as Associate Professor at McGill University’s Schulich School of Music. “In my research, I’m now experimenting with 3D surround recording, adding mics for height channels to deal with the ‘Z’ axis. We have a room with about 30 loudspeakers in it, so we’re doing all kinds of complicated multi–channel mixing now. I’m having a lot of fun with it!”
If you’d like to see some video footage of Richard King’s studio setup for The Goat Rodeo Sessions (along with some of his other recordings besides), you can find a selection of in–the–studio videos on Yo–Yo Ma’s own web site.