Wrench in the Works

Welsh mix engineer David Wrench appeals to artists with a difference. Jungle, FKA twigs, Caribou; they’ve all gravitated to Wrench to tighten up their albums, and it’s bringing the format back.

18 January 2015

Portrait: Richard Ecclestone

“I got into engineering a bit late,” says David Wrench, with some understatement. The Welshman’s unusual career trajectory has led him to become, at the tender age of 42, one of the UK’s hippest and most in-demand mixers. This year alone several albums that he mixed, most notably Jungle’s Jungle and FKA twigs’ LP1, made big splashes and have been creeping up the international charts all over the planet. Both albums have a strong electronic music vibe with an experimental edge. With the critics raving — and Jungle and FKA twigs both being young, up and coming British acts — their respective debut albums also have an immense amount of street cred.

At his garage studio in Bangor, north-west Wales, sheltered from the music industry Joneses, David Wrench muses on how he has arrived at his current enviable position, after having spent “many years trying to work out what to do with my life.” Wrench set his first tentative steps in a recording studio as a teenager when Gorwel Owen, his high school teacher — who later produced Welsh bands Super Furry Animals and Gorky’s Zygotic Mynci — took Wrench into his studio and showed him “how stuff worked.” Following this, Wrench focused on a music career of his own, singing and playing keyboards, and releasing three solo albums.

The world’s relative indifference to Wrench’s solo releases triggered some soul-searching, which was resolved in his late twenties when he landed a job as an engineer at Bryn Derwen studio, six miles from Bangor. “I didn’t really know much about engineering,” recalls Wrench, “so I taught myself. I put lots of microphones on things and found out what worked and didn’t. I worked with local bands for six months, and then did my first proper record with an amazing songwriter called Jackie Leven. The first album we did together was Defending Ancient Springs (2000) and I ended up doing 14 albums with him!” Wrench continued to work at Bryn Derwen for several more years, gaining a reputation as one of the foremost producers of Welsh-language music, which earned him the BBC Welsh radio Producer of the Year award in 2007, 2009, 2010, 2011, and 2012. He also occasionally mixed projects, and it was his mix of the award-winning Caribou album Swim (2010) in particular that earned Wrench a reputation as a top-level mixer.

FROM LEFT FIELD TO JUNGLE

Since mixing Caribou’s Swim, the amount of mix work Wrench has been doing has risen sharply. In addition to his Jungle and FKA twigs mixes, Wrench has this year also mixed Caribou’s Our Love, four songs on Owen Pallett’s In Conflict, Fanfarlo’s Let’s Go Extinct, Radiohead drummer Phil Selway’s Weatherhouse, the debut album of the British band Glass Animals Zaba, which was a near top 10 in Australia earlier this year, as well as Oz acts Seekae and Jack Ladder. The common thread on all these releases is a willingness of the artists in question to experiment and push the limit. Given that Wrench’s own artistic output is as left field as it gets — his recent album Spades & Hoes & Plows features what he calls “anti-recording” and his early noughties singles had titles like Superhorny and Fuck You And Your War on Terror — run-of-the-mill, hit-parade fodder clearly isn’t his cup of tea. Wrench himself calls his tastes “eclectic” and qualifies this by adding that he, in fact, quite likes some pop music. Nonetheless, given Wrench’s track record of working with left-field artists, he is predominantly hired for his eclectic tastes, and, one presumes, eclectic approach to mixing.

“Yes, I think this is the case,” confirms Wrench. “I’m not into polishing things and making them sound as smooth as possible. I like stuff that leaps out in the mix. I like things to be exciting. I don’t like mixes that are too tasteful. I like things that sound a tiny bit unbalanced in the mix, with things not necessarily wrong, but set at a volume or placed in a position where you would not instinctively put them, or EQ-ed brighter than you would imagine. I like things to rub a little in the mix, and I guess it is why people come to me. But the Jungle and FKA twigs material has been A-listed on Radio 1 [the UK’s mainstream, youth-orientated pop station], so it’s not so weird that it won’t get on the radio. Both artists make music that is very distinctive, even though underneath it contains pop songs. They are just very deconstructed pop songs!”

While Wrench’s musical tastes and mix approach are eclectic, his actual mix tools at his garage studio, which he calls Tŵr Yr Uncorn (Welsh for The Unicorn’s Tower), are entirely mainstream. No endless racks of esoteric hardware in his case, just a Pro Tools 11 system with an Omni interface, Adam A7X monitors with a Focal CMS5 sub, and a few bits and pieces, one of the main pieces being a Presonus Central Station. In addition, explains Wrench, “I also have Yamaha MSP5 monitors, and a Pure stereo radio that I also monitor through. I swear by my Beyer DT880 Pro headphones, and have a Radial Duplex DI, a Radial PRO RMP for reamping, and a Lehle splitter for separating signals to various amps, plus a few mics for when I go out recording. Finally, I occasionally send signals to my old Aria stereo spring reverb.”

Wrench clearly conducts virtually all his mix weirdness in-the-box, even though, he recalls, “I started working on tape at Bryn Derwen. The studio had a Soundtracks 24-track, and then changed to an Otari MX80. Learning on tape was brilliant. One of the things I’ve noticed that’s different between engineers who started on tape and those that didn’t is those that did are much more careful about gain structure, making sure it’s right all the way through. It’s still the first thing I do when I get a Pro Tools session in to mix; make sure the gain structure is right from start to end, because it gives you a lot more headroom to work with. You don’t end up with such enormous problems in digital if your gain structure is wrong, but being rigorous about it does make a difference.

“I stopped working at Bryn Derwen a few years ago because I was getting more and more mix work, and it made sense to do it in my own studio. There are no longer the budgets to spend lots of time mixing in a commercial studio, and also, being in-the-box I am entirely flexible, and can switch between projects easily. At the end of most projects I go to Strongroom studios in London for a few days, just to finish off. I like to mix with a sub, because I want to know what’s happening in the low end, particularly when working with dance music, and the overall frequency balance at Strongroom is very comparable to that in my studio, so it allows me to verify exactly what’s going on. Also, when working on the recent Caribou album Dan [Snaith] and I spent a week together at the end of the project. We would go into a nightclub every evening just before the people came in to blast the mixes through the club system and make sure it sounded great. Music is played on so many systems these days, in clubs, cars, on iPods, hi-fi stereo systems, radio, and laptop speakers, you need to put in quite a bit of effort to make sure it works on all of them. The low end of bass-driven music can disappear on small speakers, so I often end up boosting the low mids, or if they’re not there I distort the bass line to get some harmonics, then roll off the low end and mix that back in with the bass to make sure the bass line has enough definition on systems with smaller speakers.”

TWIGGING TO IT

In-the-box flexibility was of crucial importance for Wrench when mixing the Jungle and FKA twigs albums, as he mixed both, as well as Caribou’s Our Love, in the same time period. “It was a bit of a crazy time,” recalls Wrench. “I began with Jungle, and then a week later I started on twigs. The Jungle mixes were more straightforward, because they sounded good when they arrived, and the balances were right, so there were less creative decisions to be made. It was about getting their mixes sounding better, bigger and clearer. Josh [Lloyd-Watson] and Tom [McFarland] recorded and played everything themselves and did all the production themselves. I came into it as an outside person, which meant the main challenge in the beginning was to establish trust. I did a trial mix for the album, in fact I did the same for the FKA twigs album. I mixed just one track in each case to show them the direction I would take. In both instances I apparently got it right and finished the entire album.”

According to Wrench, mixing the FKA twigs album was far more difficult, in part because of the plethora of co-writers and producers involved, including Emile Haynie, Arca, Joel Compass, Clams Casino and the ubiquitous Paul Epworth. Wrench: “The sessions came in all sorts of different sample rates, formats, levels of organisation, and musical directions. There was a unifying element, but it wasn’t obvious to start with. Part of the challenge was to get all the tracks to sit together as an album, and yet make sure they all remained distinctive. Twigs has really strong ideas of what she wants things to sound like. She knows every single detail of each song, even things that are really buried in the mix or even muted, and where she wants them placed in the mix. Much of it has to do with the way she dances, and how she feels the rhythm and the movement of the song. She does a lot of the drum programming herself, using the Dave Smith Tempest, and she works very quickly and instinctively, which gives her music a kind of confrontational punk feel, even though it’s mostly programmed. This was another reason why quite a few things needed sorting out in the mix, with me putting things in some sort of order, while also keeping the edge and energy that typifies her music.

“Organising a session is generally the first thing I do when I get a mix in. I go through it and make sure it’s something I can work with, putting things in the right order, getting rid of clicks and other noises, making sure the gain structure is OK, and so on. I then work very quickly and really instinctively, however big the session is. I like to go through and get a basic balance as fast as I can. I tend to first work on the drums and bass, or whatever is the main body of the music, and then I get to the main chordal information, and then the lead vocals, and finally all the effects and interesting little bits and pieces, and the backing vocals. So I initially get the track to the point where I have a feel for it, where it excites me and moves me, and after that I go into all the details. I also always keep the rough mix at the top of the session, and send that out to a different pair of outputs, using the Presonus Central Station to be able to flick back and forth between the rough and what I’m doing, and reference the rough, allowing me to check whether I’m improving on it, and not making it worse.”

ALBUM ON TRACK

Time is the fourth single from Jungle’s debut album, released in September 2014. It’s the highest-charting Jungle single, at a meagre 94 in the UK. FKA twigs’ highest charting single in the UK, Two Weeks, only reached number 200, meaning both Jungle and FKA twigs are newly arrived exponents of a beast that’s been all but declared extinct: the album. Wrench agrees that the success of both albums without any high-charting singles is “very unusual,” while at the same time pointing out he mixes many complete albums these days, which, he says, “makes sense when the label wants a distinctive, unifying sound across an entire album. It also makes economic sense to have one mixer do an entire album, rather than have a different mixer for each track.”

In other words, the album may have a future after all, and the Jungle and FKA twigs albums are living proof. Many reviewers have commented on the cohesiveness and singularity of vision of both the FKA twigs and Jungle albums. It’s quite an achievement, and has a lot to do with Wrench tightening the screws on these debuts.

I don’t like mixes that are too tasteful. I like things to rub a little in the mix, and I guess it is why people come to me

01_session-screenshot-for-Jungle’s-Time-pichi

02_session-screenshot-for-Jungle’s-Time-pichi — The session screenshot for Jungle’s Time including layers of musique-concrète-like samples; and a typical vocal chain on the project.

JUNGLE’S TIME SESSION

The music on Jungle fits loosely in the neo-soul category, featuring strong influences from 1970s soul and disco music, updated for the 21st century, with big bass, shiny high end, falsetto vocals, and multi-layered musique concrète-like real life samples adding to the excitement.

Wrench recalls, “When I first met Tom and Josh we talked about music a lot and we connected over our mutual like of Shuggie Otis. But they also listen to bands like Can and other things, so they incorporate a great variety of influences. They have a very defined idea of how they want their music to sound. My main concern in the beginning was that their rough mixes were a bit muddy, and I made things too bright and clear on Busy Earnin’, the first track I mixed. So they asked me to pull that back a bit again. A lot of the bass on the album has a chorus effect on it, so it had a kind of watery warmth, and sounded a bit as if someone had thrown a blanket over the mixes. For me it was a matter of getting more clarity and punch, while retaining that warmth.

“The song Time was the last mix I did for the album, which made it a bit easier, because by then we had worked out the general approach. This meant the sessions I received were also closer to what the finished mixes would sound like. After I mixed Time they felt something was still missing, so they went back to their studio and added a few additional bits and pieces for it, like the guitar that enters in the second verse, and there’s a bit of extra excitement at the end, like the synthesiser flourishes. The main thing about this track was to get it to sound punchy with driving drums, and to get the vocals sounding clear, using automation to make sure things are leaping out and exciting in certain places.

“In fact, almost all the vocals on the album went through a Leslie effect, even though I would occasionally add in a clean vocal for a bit more clarity. But the Leslie effect on the vocals was the sound it had, just like they’d stuck the bass through a chorus. The Jungle guys work in Logic, so they’d stem out their Logic sessions for me, which I imported into Pro Tools. Normally I take plug-ins and effects off when I start mixing, but because so many of their effects were intrinsic to their sound, I preferred to keep their effects. For me to have spent time to try to recreate their effects would have been daft. If it sounds good, it sounds good. So they rendered me both dry and wet stems, and I would sometimes go back and ask for entirely wet stems. In a couple of cases I might have sent a clean vocal stem back to them after I had taken out clicks and things, so they could run it through their effect chains again.”

The Time session is nothing if not intricate, with 70-odd audio tracks. From top to bottom there’s the rough mix, drums and percussion (including things like ‘sandcrunch,’ ‘coins,’ ‘bath tub’), bass and incidental sounds (named things like ‘restaurant,’ ‘crackle’ ‘gym FX’), guitars, synths, organs, pianos, trumpet, more guitars, and vocals, and finally the master mix track.

Wrench explained: “All the drums in the session are programmed by Josh and Tom, using their own samples and sounds, just like with the synths and all the incidental sounds. It’s almost like musique concrète in places! They have many background noises going on that are not immediately obvious, but if you take them out, the tracks would not sound anywhere near as exciting. There’s a track with record crackle that just adds a bit of atmosphere in the background that you don’t notice, but when it’s not there the song sounds dry. The Gym FX tracks are people playing basketball in a gym hall, you can hear the squeaks of the shoes and the ball bouncing. All these sounds create an atmosphere and subliminal narrative to the song.

“I had very few effects on the drums, just the Pro Tools EQ3 on the kick and snare, and Waves Puigchild 670 and Valhalla Vintage reverb on the toms, a room just to make sure it’s not entirely dry and give it a bit of a stereo feel. The Valhalla is my main reverb. I also use the Pro Tools EQ3 a lot. Most of my EQ is subtractive, doing little cuts here and there of frequencies that I find conflicting or that are clogging things up, and I like the fact the EQ3 is not particularly characterful, which makes it ideal to do these cuts with. I have the Waves CLA 1176 compressor on the bass, and then the Puigchild 670 on the extra guitar, which was added later on, because it needed some more character. The Prophet synth also has the 670, as well as a Mondomod. Two of the organ tracks have the S1 imager for some more width.

“The balance between Josh and Tom’s vocals is very delicate and the key to mixing them and getting them both to sit right together. One of them sings falsetto, the other doesn’t, and they have quite different voices. If one leads too much, they both sound wrong. They wanted their vocals to be quite dry. We tried reverb on a few tracks, and it does appear here and there, but the Leslie effect on their vocals is the key to their sound. I grouped all vocals to Aux 4 on which I had the C4 multi-band compressor. I am a big fan of multi-band compression. If you EQ high end out of a vocal in one place because it sounds too harsh, it can sound dull in other places, and I find that multiband compression can really control that. In the case of this song the C4 would have been there to control the high and low mids. The vocals were either getting a bit boomy or a bit aggressive, and the C4 just controls the bits where it gets too much and pulls them down. I also had the Puigchild 670 on each of the vocals tracks, for more overall compression.

03_session-screenshot-for-Jungle’s-Time-pichi

04_session-screenshot-for-Jungle’s-Time-pichi

FKA TWIGS’ TWO WEEKS SESSION

The main emphasis of the 80-odd track Two Weeks session is the extraordinarily large vocal section, consisting of 40 audio tracks. The fact that many of the vocal tracks are almost black because of numerous edits indicates how much detailed work went into the vocal arrangement and mix.

Wrench had the lowdown: “I mixed Two Weeks as part of the whole body of the album, but we always knew it was going to be the main single, so there was a bit of extra pressure. It’s also a big track, and needed lots of work. With twigs, mixing took probably one and a half to two days per song, because all the sessions were very complicated. There was a lot of cleaning up to do on these tracks before I could begin mixing, in part because of all the different producers she’d been working with. One of the briefs I got from twigs when these sessions came to me was she wanted the beats to sound really hard, with a hip-hop low-end and R&B crunch and bite. She didn’t want things to sound soft at all. She also wanted the vocals to be proper pop vocals, right up there, with the harmonies, with an R&B slickness to them. I don’t think there was ever any danger of her sounding bland, because her performance on it is so good and the melodies and the words are so arresting. So I didn’t worry about blanding-out the vocal.

“The first thing I did on Two Weeks, after getting a vague balance, was to really clean up the vocals. Some of them were quite compressed, with very loud breaths and lots of clicks, and because there were so many vocals, it felt quite messy. So I cleaned each vocal track up, and because the vocals had to be really slick and tight, I then made sure all the backing vocals were absolutely in time with each other. All the black lines show where I cut the wave forms and shifted them around by hand to get them to be really tight. I was shifting them around as if they were samples, so they were absolutely spot on. I can never get Vocalign to work properly, but I did what that’s supposed to do, just by hand. Her main vocal doesn’t have many cuts in it because the timing was great.

“The session came in at 96k and remained at that. I cleaned it up, changed the structure so it was easier for me to work with and stripped it of most plug-ins. At the top is the rough mix, as usual for me, and immediately below it is my master track. I don’t have a definite place where I put it, but it usually is at the top or the bottom. Below that are the drums, with most of them coming from twig’s Tempest. An amazing producer called Arca also worked on the track, well as Emile Haynie, so I don’t know who did what, but I do know that twig did a large part of the production on this song.

“The 808 drop is a big, heavy sound that appears only four times, but that is quite a big feature in a club, where it really hits you in the stomach. I had the Pro Tools D3 CL compressor/limiter on that, and a very sharp low-pass filter at 400Hz. Beat 1 has a snare and a bass drum, and because of this I’m using the C4 multiband compressor to control the level of the snare. The Tempest only has stereo out, so all twigs’ drum tracks came as stereo pairs. She gets the craziest sounds out of that machine; she’s really mastered it. Beat Sub is an effect channel with the Spring Reverb and an EchoBoy eighth-note delay, only 10-20% wet, and I use it on the claps and several of the more incidental samples below, like FX Scream and Sonorhits. Below that are the Toms, and then the SnareHelp, which thickens up the snare from the Beat 1 track. Beat 2 has a bassdrum-like sound on which I put a strong EchoBoy repeat echo, which you can clearly hear in the track. Like the Choruskick below, it has a spring reverb, though the Choruskick is not really a kick, it functions more like a percussion element. Reverses is a reverse snare, and the Sonor sounds are like submarine sounds.

“Below the drums are several atmospheric samples, and then the main bass track, which has a C4 to keep it in place, and below that six yellow-coloured synth tracks, that also operate in the low end and all have a similar sound. The MainSynth has an S1 Imager to widen it, and the MainSynth 2 has a Microshift, which is automated and comes in at certain sections. I also have a Mondomod on one of the ChorusSynths, again to widen it. You can also see that I have volume automation on another ChorusSynth, increasing its level in each chorus to help the choruses build. Further down are five green Bridge Synth tracks, and it was quite tricky to get them to sit right, as they are all panning around a bit, which is why I have two instances of the Mondomod going on, to help with the width and the sense of movement these parts needed. The Microshift plug-in does some widening and some delay and pitch shifting. Again you can see how the volume increases during the bridge section, helping to build tension.

“The vocals start with the pink coloured verse backing vocal tracks, which had quite a lot of background noise, which is why they are chopped so much between each line. VerseLV11Fb is the verse lead vocal, which is quite natural. I brought some of the breaths down, lifted the occasional word and compressed it quite a bit, using the PuigChild 660. There is also automated EQ from the EQ3, coming in on some words only. This is also why there’s a DeEsser, which is essential if you use serious compression — I use the Waves Renaissance DeEsser. VBVT11F01 has just one word on it, ‘high,’ which has a big, long reverb from the ValhallaShimmer and an EchoBoy delay. All this send’s bus tracks are automated and come in for different sections. There’s a hell of a lot of automation on that vocal!

“BrdAdLbC1, aka Bridge Ad Lib, contains just one word, ‘hah’, and has a whole bunch of plug-ins on it, seven inserts and two sends. I think there originally was more on this track, but in the end we whittled it down to just one word. Below it are several other bridge vocals, again with loads of automation and again cut heavily and moved around, to make them sound like one instrument, and sent to the Bridgebus aux track below the two Lowbridge vocals, with an EQ3 and the PuigChild 670. The two Lowbridge tracks contain just a phrase at the end of the bridge, which is pitch shifted and has spring reverb. Then there are four more Bridge vocal tracks, which actually sound in the chorus, and go to the Aux 4, on which I have the EQ3, Waves CLA 1176, Renaissance DeEsser, another EQ3, and an Echoboy.

“Below the Aux 4 track are the chorus lead vocals, which go to the Tape LV Bus aux track, on which I have the SSL EQ, Waves Q4 EQ and the Renaissance DeEsser. Below this are two chorus backing vocals, which go to the Chorus Stack Bus Aux track, which has the same plug-ins on the insert as the lead vocals, plus an Echoboy delay, and five sends! There are also two BridgeAdLibCT tracks, which contain the phrase ‘Higher than a motherfucker’ in the chorus, which we had to remove for the clean version. For BBC Radio 1 we couldn’t even use the word ‘thighs’! This line again has two instances of the EQ3 7-band, the Renaissance DeEsser, and the PuigChild 670 compressor, plus several buses on the sends, and further down are the Hivoxbounce and OctDPrint tracks, which are just a few words in the choruses, but have tons of effects on the Sends, which result in long tails that make them sound quite angelic. There’s a LowOct track which you can hear right at the beginning of the song, which is pitch shifted and sounds like an underwater vocal. The EnOvPrint track has a PanMan for a kind of left-right tremolo effect, sounding right at the end, and adding another layer.

Two Worlds has an extraordinarily complicated vocal mix, and Wrench laughed when it was suggested he may have set a world record for the amount of per-word effects. “You’re right, it really is a very complicated vocal arrangement and because of the way I treated it, a very complicated mix as well! The whole album was like this. And twigs knew every single detail and how she wanted to hear it. It was really inspiring to work with someone like that.”

“I always have similar plug-ins on the master track. I first go through an SSL EQ, for some very gentle EQ, and then through the SSL bus compressor (I used to use the Smart C1 when I still used outboard), and I will have that set on a slow-ish attack and a fast release and a 2:1 ratio. I will often automate the make-up gain on that for various sections of the song, because after that it goes into the Linear Multiband compressor, and I will drive the choruses of the song a bit harder than the other bits. It may just be 1-2dB, but lifting the song in some sections is part of creating excitement. Then there’s another EQ3, just touching things very slightly.”

REVIEWED