Session Notes: Elephant Tree

Riley looks on and listens from behind The Church Studio 1’s EMI Neve console, while his Elephant Tree bandmates play through the track. A key benefit of the open-plan layout is the ease of communication between the engineer and performers.

We make a rare visit to a world-class recording studio, as one of The Church’s engineers takes us through a recent session.

I’m in the happy position of working as an in-house Assistant Engineer at Paul Epworth’s The Church Studios in London. [See our video of the studio at http://sosm.ag/paul-epworth-the-church.] This isn’t usually the kind of facility we see being used in Session Notes, but when SOS invited me to write a feature based on my recent recording session there with the band Elephant Tree (in which I play, see box), I thought it would be a great opportunity to explain what it’s like to record in such a studio. If you’re used to working in more modest surroundings, you might be surprised at just how similar the techniques and working practices can be in such contrasting environments.

Room For Doom

The one-day session (I also had the following day for mixdown) was in Studio 1, a large room on the second-floor of this converted 19th-century church. It features a vintage 72-channel EMI Neve console, and a host of modern and vintage outboard gear, synths, instruments, amps and other toys for musicians and sound geeks. There’s also an Endless Analogue CLASP system, rigged up to a Studer A80 Mark IV two-inch analogue tape machine. We used that on this session, partly just because the opportunity was there and partly because I knew from experience that analogue tape usually suited the band’s sound.

Although Studio 1 is a great space in which to work, it posed some questions. Recordings in the doom and stoner-metal genres have a fairly defined aesthetic, one aspect of which is a very obviously lo-fi sound. (I generalise, of course, but those who have produced higher-budget records are the exceptions that prove the rule.) While I planned to draw on elements of the doom and stoner aesthetics and to track the band playing together, I decided to see what would result from recording things in a more conventionally ‘good’ way in a nice acoustic space, as I could see some ways in which this particular space could be used to advantage. I was particularly excited about the possibility of using it to achieve a great doom-metal drum sound. We had a very distant and roomy-sounding kit that feels natural but still ‘big’ enough to sit with the heavy guitars the genre demands. I figured I’d be able to achieve this using ambient and room mics in this large room, and would avoid an over-reliance on artificial reverb.

Some other aesthetic elements typical of the genre include loud, detuned, fuzzed-out and crackly distorted guitar sounds, with frequent, unabashed use of delay and phasing. Equally distorted are the fuzzy bass sounds used to create a thick, dense low-end, which often remains fairly indistinguishable from that of the guitars. Vocals are commonly sung clean, drenched in effects, and sit deep inside the mix, although some bands, including this one, use shouting vocals as well.

Hall Monitor

While the Neve Console is arranged in a traditional way, in that it faces out onto the tracking floor with outboard gear in racks behind it, Studio 1 has no separate control room. There are diffusion panels on the wall behind the console and an acoustic ‘cloud’ hangs above, but the desk sits in the same room as everything else. The ‘control area’ occupies about a quarter of the space. While home or location recordists may be accustomed to monitoring in the same space as musicians play, it’s far less common in pro recording studios.

There are obvious disadvantages to an open-plan studio, most notably with monitoring, but I believe the advantages in sessions like this outweigh them. The ease of communication when there’s no barrier between engineer and band makes the whole recording experience feel more comfortable and less alien. There’s no need for talkback mics, as the band can just take off their cans after a take, and we can chat at a regular speaking volume. It’s also easy for the engineer to walk out and adjust a mic, work on the tone and effects parameters with the guitar player, or catch the attention of one of the players and make suggestions while the band are playing.

Splendid Isolation

Given the genre, that it’s typically played very loud, and that the guitar player uses lots of feedback, an obvious option was to keep all the players, with their instruments and amps, in the same room — this often makes for a better performance (it’s hard for bands to vibe on heavy music when they can’t feel the ‘weight’ of their amps at their back), I knew I’d be mixing the track, and I’m not terrified of working with bleed! Nonetheless, I chose to isolate the guitar and bass cabs, for several reasons. For one thing, these musicians had been in this studio before and so were comfortable with the setting and the Aviom headphone monitoring system.

This system would allow me give each performer a separate mix that they could balance against their own part to taste and crank up as loud as they liked. I also knew we’d be overdubbing more guitars, and figured that would be easier to do with the amps in the same room as the guitarist. Another consideration was that, as the band would be developing the track’s arrangement on the day of recording, it was possible that we’d want to alter the arrangement after some takes had been captured; isolation would be helpful in the event that edits were required. Finally, making subtle adjustments while listening on headphones is enough of a compromise with the drums playing 20 feet away, never mind throwing loud guitar and bass into the mix as well!

Caving In

Here you can see some of the close mics used on the drums, including a Neumann U47 FET and Yamaha NS10 driver on the kick. Riley took time to reappraise some of the mic choices after the band had played a little.A drum ‘cave’ was erected, largely to help keep the sound close and tight, with the main image being captured not by the overheads, but by two pairs of room mics.I wanted reasonably ‘tight’ sounds from the drum close-mics, as I’d use these mostly to supplement the room and ambient mics if I felt they lacked anything. For this reason, I erected a sort of ‘drum cave’ in the centre of the room, using seven 2.5m-high baffles. On the kick I initially placed an AKG C414 B-ULS large-diaphragm condenser mic, set to cardioid pattern, just outside the hole on the resonant head to pick up slap and shell tones. However, I eventually replaced this with a Neumann U47 FET, a classic mic choice for kick drums that just sounded better to my ears. A pop filter served to reduce the air pressure at the mic’s diaphragm without really detracting from the sound. I also used a Yamaha NS10 speaker driver to augment the kick’s low-end with a nice, meaty ‘woof’. I opted for Josephson e22S side-address cardioid condenser mics on the two toms, as well as the snare top to start with. The side-address configuration makes these mics great for drum close-miking, as they’re really easy to place.

A spaced pair of Neumann TLM 170 cardioid capacitor mics took care of stereo-overhead duties. I’d typically use ribbons on top of the kit for this kind of music but I wanted these overheads to add brightness to the cymbals if it were lacking in the room and ambient mics. I kept both mics fairly low to keep things sounding tight but, to avoid phase-cancellation problems, I placed one slightly lower than the other, making them equidistant from the snare. The left one (audience perspective) was somewhere between the crash, ride and floor tom, and the right between the hi-hat and rack tom. I also placed an old AKG D19 as a mono central overhead, the idea being that this would give me the option of a little more ‘crunch’ when mixing.

When recording, I usually think of the stereo image of a drum kit as being arranged around a line running through the centre of the snare and the centre of the kick’s resonant head, which means room-mic configurations must be slightly ‘skewed’ to one side. As drummer Sam was using very heavy, dark-sounding cymbals, I wasn’t worried about the kit sounding too bright or harsh so I first tried a Neumann KU100, a binaural ‘dummy head’ mic, for this job — it’s a mic that can sound great on lots of things but is sometimes rather cold and clinical. I placed this about five feet away from the kit, with the kit at the centre of the stereo image and the snare right in the middle of the kit. I figured this mic might help add ‘crunch’ to the sound if anything, but I soon discovered that wasn’t giving me what I wanted so, after several abortive experiments to warm things up using some nice vintage hardware compressors, I decided to cut my losses and try something completely different: a pair of Coles 4038 ribbon mics. I set these about six feet back from the cave, just outside the edges of the hi-hat and floor tom, pointed slightly inward, with the snare and kick at the centre of the image. This sounded much better! Further back in the room, just in front of the console, were two B&K 4011 omni-directional condensers, mounted on tall stands (about 10 feet up), spaced evenly about eight feet apart and 15 feet away from the kit.

Signal Chain

All the drum mics were plugged into tie lines normalled to the console’s Neve 1091 mic preamps. To understand the way I used the console, you need to appreciate that these EMI Neve channels work in three modes, called Mon, Track, and Group. These effectively assign the destination of the 24-button routing matrix at the top of the channels, the level being controlled by the fader on that channel. In Mon mode, channels 1 and 2 are routed to the main mix bus, and 3 and 4 to the ‘back’ bus. In Track mode, channels 1-24 are routed to the corresponding multitrack-recording sends. In Group mode, channels 1-12 are routed to the group outputs.

When using the CLASP system (reviewed in SOS May 2012: http://sosm.ag/ea-clasp) the musicians have to monitor directly off the input channels in order to hear themselves before the tape machine delays the signal en route to Pro Tools. As this console only has four pre-fader sends and because I’d like to be able to send more than four separate instruments to the headphone pods, I chose to run the channels in Group mode, and patch out of the post-EQ insert send of the channel and straight into the multitrack send (after passing through any inserts). That way, I was able to use the 12 group outputs to send whatever I wanted to the headphones. Normally, running it this way would mean that the mic/line gain pot is the last port of call for adjusting level before hitting the tape (unless there happen to be any inserts with gain controls on any given channel). However, built into this particular console is a post-mic and line amp and a pre-EQ attenuation pot — when sending to the multitrack from the post-EQ insert send, you have the ability to attenuate the signal going to the tape. As the 1091’s mic gain controls increase gain in 5dB steps, this feature is really useful for fine-tuning the level. It always pays to learn the intricacies of a console you’re going to use regularly!

A couple of The Church’s vintage EMI toys that were used on the drum room mics: a pair of EMI TG12413 Limiters for the Coles ribbons that provided the main room sound, and a Zener Limiter for the ambient B&K omni mics.I decided to use very little outboard processing and EQ on the drums while recording. One reason is that I knew that I’d be mixing this track in the same room the following day. I’d thus have access to all the same gear but, importantly, also to the monitor speakers, rather than having to rely on headphones. I also knew I wanted to do most of the compression in parallel, which is easier with many outboard compressors when mixing. Still, I did use two 500-series Elysia Nvelopes (a kind of transient shaper: http://sosm.ag/elysia-nvelope500) on the close kick and snare mics, so I could dial in some extra punch and crack if needed. I also used an original EMI Zener limiter on the ambient mics to enhance and exaggerate the sound of the room — I knew I’d be doing this anyway, and it just made sense to commit to this decision rather than postpone it. For similar reasons, I deployed a pair of EMI TG12413 Limiters on the Coles room mics.

Bass: DI & Fuzz

As the Electro-Harmonix Big Muff Pi pedal, which was responsible for the bass fuzz sound, sucks out some bottom end, Riley used a Pultec EQ to apply a hefty LF boost to the bass part while tracking.The bass guitar rig was set up very simply. I used a Tab V71 DI box, taking a clean DI signal, before passing it through the bassist’s stompboxes and out to a modern Ampeg B-15 Fliptop amp. Placing the bass cab in a separate booth at the back of the room, I miked it with a Neumann U47 FET condenser, which I placed on axis about four inches from the grill, and three inches to the side of the dust cap. The DI and U47 signals were run down tie lines and patched straight into the console.

I also applied some compression and EQ while tracking the bass. Again, that could have been left until later, but why defer work you know must be done? I compressed the clean DI signal quite assertively through an Empirical Labs EL8-X Distressor so that it was knocking off up to 10dB at the loudest parts. The level of the distorted signal is naturally more static than the clean DI one and I knew I’d end up blending the two (both to contribute to the overall tone and to supplement the low end in particular), so I wanted to reduce the variation in level difference between the two tracks. Compression also meant that I could better control the level going to tape, allowing me to hit the tape a bit harder for effect without risking complete overloading. I didn’t want a pumping, or audibly heavy compression effect, though. I opted for a reasonably fast attack time, to knock down some of the big peaks, and a fairly slow release, set to try and make things sound fairly natural. The playing was fairly dynamic and despite the 10dB headline gain-reduction figure, it probably averaged something closer to 5-6dB.

An Electro-Harmonix Big Muff Pi makes a key contribution to Pete’s primary bass tone, which is heavily distorted, so I didn’t want to compress the miked-cab signal too much more. Still, I passed it through an EAR 660 valve compressor but set for only three or four decibels of gain reduction at most. Also, because the Big Muff Pi causes some fairly significant low-end loss, I used a Pultec EQP-1S EQ after the compressor to add a healthy low-end boost. Later discovering that I had a spare track left on the tape, I decided to send some of the bass DI signal through a Thermionic Culture Vulture tube-distortion processor and record the result, via the console, to the spare track — this would give me yet another tone with which to finesse the sound while mixing, without using up precious time patching in processors or making irreversible decisions on the fly.

Axe Tracks

Two amps were used on the electric guitar during the main band session. Both were miked in similar fashion, with a Shure SM57 alongside a Royer R121, and placed in isolation booths. The Marshall rig pictured here was also used in the main live space for overdubs, where it was joined by a slightly more distant Coles 4038 mic to fill out the sound.For the electric guitar, I chose to capture a clean signal via a DI box before going through any of the stompboxes. I always capture a clean DI feed when recording distorted guitars, partly because the visual waveform is rather easier to work with as a reference, should editing be required, but also as it makes re-amping an option which, given the time constraints, I felt would be useful. At this stage in proceedings, given that the band needed to figure out an arrangement for the tune, getting the band up and running, happy with their headphone mixes, and pulling together a decent drum sound were my main priorities. Having the ability to re-amp meant that I could afford to neglect the subtleties of the guitar tones a little bit if necessary for the live takes. As we’d be doing overdubs anyway, re-amping wouldn’t be the hassle it sometimes can be.

After the DI, the guitar signal went through guitarist Jack’s stompboxes and was then fed down an XLR cable into a Little Labs STD Line Driver, which split out into a Marshall JMP Head with a 4x12 Marshall cab and a Fender Twin Deluxe Reverb 2x12 combo, that were placed next to each other in another isolated booth. I used what’s as near to a standard pair of mics on each amp as you can get: a Shure SM57 dynamic mic and a Royer 121 ribbon. It’s nice to search for unique sounds sometimes, but I wanted something tried-and-tested that would give me a good-sounding guitar tone quickly and experience tells me I’d probably have ended up with this combination for heavily distorted guitar in any case, even if I’d had all day to get a sound! Both mics were placed a few inches away from the grille cloth, on both amps. For the Marshall I started with both mics on axis in a coincident-pair configuration, placed roughly halfway between the dust cap and the outer rim of the speaker (always make the effort to listen to the sound of the different speakers in a cab — one usually ‘sings’ a little more sweetly than the others). On the Twin, I used a broadly similar technique, though this time placing them slightly closer to the dust cap.

Having a nice range of outboard preamps at my disposal, I decided to plumb all four guitar mics into Chandler EMI TG2s, patching the outputs of those into the console’s line inputs. After some minimal shuffling of the mics and flipping of signal polarities (‘phase’), I was happy enough with the sounds to move on. I kept the guitar mics on separate tracks at this point, rather than submixing them, because I had enough tracks available on the tape. Having worked quickly on these sounds, I preferred not to have to commit them to a mix at this stage.

Last-minute Arrangements

I now had the band up and running, headphones working, and everything looking good in the Pro Tools session. I was happy with what I was hearing in my headphones too so I started recording the passes the band were making as they continued to work out the song. When we seemed close to arriving at something we were happy with structurally, we began taking breaks to listen back and appraise things, and I was able to make some compression and EQ tweaks, put in more work to refine the phase relationships between drum mics, and hone in on a few other issues.

I was getting a lot of bleed from the hi-hat into the snare’s top mic. I wasn’t overly thrilled with the snare sound either, so I decided to try a different mic. A Shure SM57 seemed slightly better but still not stellar, so I swapped that out for a Beyerdynamic M201. This dynamic mic has a hypercardioid polar pattern, which is tighter than that of the other two. In the past I’ve A/B-compared M201s and SM57s on snares and not always found the difference to be dramatic, but in this case it really did the trick, both in terms of the snare sound itself and rejection of the hi-hat. I was also slightly unhappy with the under-snare mic, which sounded a bit harsh and brittle to my ears. An SM57 smoothed out the sound nicely. The other mics all sounded pretty good so, after we’d nailed down the final idea for the arrangement, we had a brief stop for lunch at the pub to recharge. Then we began laying down some full takes.

I’ve recorded my own bands a lot over the years and used to prefer to go into recording sessions with arrangements that had been rehearsed extensively and usually played live at least a few times. More recently, however, I’ve come to see some of the benefits of going into the studio with just a rough idea, or even nothing at all. It’s not a luxury many people can afford today — at least not in this sort of facility — but it’s nice to have the chance to capture the energy and excitement around working out a new song (or in this case, a new arrangement of an existing song). A risk, though, is that when tracking this way the focus becomes playing the song correctly rather than playing it well. I find this happens often and it requires you to be particularly vigilant when you’re both engineer and band member. You need to avoid getting too excited just because the band made it through the whole piece without a bum note or a bad drum fill! To this end, I made a deliberate effort not to listen too hard for mistakes but, instead, while still paying proper attention to the technical aspects of the recording, I trained my ears on bigger-picture issues, like tempo, section changes, and the overall feel and energy in the performance. (In other words, the sort of thing most listeners care about!) A badly misplayed part was unlikely to go unnoticed by all of us anyway, and working in this way it took only a few takes to capture a performance that we all liked.

Overdubs

With the core of the song down, and everyone feeling good about it, we moved on to the guitar overdubs. I brought the Marshall stack (my favourite of the two amps) out into the live room and stuck with the SM57/R121 setup that seemed to be working, although I took time to experiment a little more with their position. The SM57 ended up slightly further from the grille cloth, aligned with the dust cap but set off axis and pointing out into the middle of the speaker cone. The R121 was placed beside the SM57, on axis, directly in front of the point on the cone that I’d aimed the SM57 towards. I played with the phase relationship of these two mics (by moving them around a bit) and arrived at a slight comb-filtering effect that I liked the sound of, so I left them there. That said, a slight hollowness in the sound that resulted from the comb-filtering bothered me. To counter this, I added a Coles 4038 about three feet in front of the cabinet, pointed right at the middle, between all the speakers. Happily, this sounded great right off the bat, serving to thicken the sound somewhat. I also decided to leave the B&K stereo room mics where they were, just in case they’d add any useful ambience.

With the mics in place, Jack stood in front of the amp as he would at a gig and turned it up — very loud! After tweaking the preamp gains accordingly, we got a solid pass of the rhythm guitar track pretty quickly and with a decent guide track in place it was easy to punch in to correct a few mistakes. The overdubbed guitar completely changed the energy of the track in a good way: things were really starting to sound big and heavy. We decided to do another pass of the slow, heavy middle section using a low-octave effect courtesy of Jack’s EHX Micro Synth pedal, and that really helped give the middle section the sort of sludgy, laborious feel we were aiming for.

We’d not yet decided as a band if we wanted a guitar solo, but had left plenty of time in the arrangement to have one in the middle section so we decided to get a lead tone going and give it a shot. It wasn’t happening! Jack wasn’t feeling it and I wasn’t digging the tone in context either, so with time ticking we decided to park that idea and capture some vocals.

Voice Of Doom

These guys like to record their vocals very quickly. They’re good enough singers, and it doesn’t take long for them to get decent takes, but it’s difficult enough to get a double track out of them, much less enough takes to do a comp or anything like that. That’s punk rock for you! With this in mind, my plan was that we’d each use the same mic to speed things along, so I set up a single Neumann U47 mic (the classic tube version this time) in the middle of the room, so that whoever was singing could easily see me behind the desk. The mic fed a Neve 1091 Mic preamp, which was routed to a Urei 1176 FET compressor-limiter. I used the latter to compress things on the way in for similar reasons as I did on the bass, though this time I didn’t track to tape but directly to Pro Tools. I opted for a slowish attack and medium release setting so that the vocal part sort of ‘rode into’ the compression, rather than making it sound ‘stepped on’, despite up to 7dB being knocked off the louder syllables and more stridently sung lines. From there, the signal was delivered to a 500-series Maag EQ, a desk-style EQ with a lovely ‘air’ band shelf boost. I didn’t really use this in the end but it wasn’t detracting from the sound in any way, making it a potentially useful device to have in the chain in the event that any tweaks were required for one singer or another.

I also used an SE Electronics Reflexion Filter behind the mic, just to tame the room sound down a bit. That was a kind of insurance policy, really: I knew the vocals were going to be very wet and very deep in the mix, but it’s hard to remove room sound after the event, and by this stage in proceedings I was more concerned with getting a chance to record them at all than with getting a truly amazing vocal sound! I could always add more ambience later.

The vocal recording process proved a bit chaotic with Jack, Pete and me switching out frequently to try different ideas, melodies, and harmonies. In such situations, an engineer needs to inject a little method into the madness so, all the while, I listened to and tweaked the levels for each singer. By the time we were tracking in earnest I had recall settings written down for each vocalist, so I could set up very quickly for the best results. Eventually, we found both a tone and a melody we all liked, and I even persuaded Jack (who ended up providing the lead vocal) to do some double tracks, which sounded great. We couldn’t have spent more than two hours on vocals for the whole tune, which mostly comprised separately tracked three-part harmonies. That’s not to say more time would have hurt but I felt it had gone well.

Sans Solo?

The unique-sounding distortion on the lead guitar part was the result of some quick experimentation with this Digitana recreation of a rare EMI Synthi HiFli effect — something that helped inspire the performance of a part that had initially seemed problematic.After the vocal success, we were ready to give the guitar solo another go and with the vocals in place it was much clearer to us what was needed. We were looking for a really mean, dirty and broken-sounding lead tone that would play something quite dissonant through the heavy middle section... and I knew immediately what I’d try first. The Church has a recreated EMS HiFli guitar synth, built by theoretical physicist and synth wizard Dr Steve Thomas of Digitana Electronics. The HiFli can create a wide range of interesting tones and modulations. Most importantly for our purposes, when driven hard it can produce incredible, unique-sounding distortion and fuzz effects. We drove the input just a little bit harder with a bespoke germanium overdrive pedal, also built by Steve, and patched Jack’s EHX Micro Synth after the HiFli to customise the tone even further. It was all very experimental and enormous fun. We ended up with a bitty, grainy, monstrously filthy lead-guitar sound that really pleased us and, with that in place, it didn’t take Jack more than a couple of passes to put a suitably grimy solo in the bag.

Eastern Promise

All that was left was to record was my sitar, an instrument I’ve been playing for several years, yet which I’ve only had the opportunity to record seriously a few times. In some respects it’s not entirely dissimilar from many Western stringed instruments, but in others it seems completely alien. It can produce an incredible amount of harmonics which somehow pour off it from everywhere and in every direction! While the bright overtone harmonics that ring out from the sympathetic strings can sound very lush and beautiful, some playing styles can make it sound harsh and sometimes almost banjo-like. This particular sitar was a nice instrument, but the strings hadn’t been changed in a long time and it was definitely capable of sounding harsh! For this reason, and because I wanted to capture a nice roomy sound, I decided to mic it using a Sontronics Apollo, which is a phantom-powered stereo ribbon mic in Blumlein configuration. I also placed a Neumann KM86 omni mic about 15 inches in front of the top of the headstock, as this is where many of the brightest and most lush-sounding harmonics seemed to emanate from. Another aim was to place the second mic far enough away from where the metallic plectrum (called a ‘mizrab’) strikes the strings, to try and capture more of the resonant sounds than the direct sound of the strummed strings. This tactic worked pretty well — the Apollo sounded warm and roomy, which was great, and I was able to mix in the KM86’s brightness to taste. I really should experiment more with recording this fascinating instrument!