Inside Track: Shawn Mendes 'Treat You Better'

George Seara.

Is mixing on a large analogue desk still worth the trouble? For George Seara and Shawn Mendes, it seems to be working out quite well...

In the last few years, Toronto has gone from being a music-industry backwater to one of the hottest sources of new talent. The Canadian city and its surrounding areas have given us artists as diverse as the Weeknd, Drake, Magic! and Justin Bieber, and the latest flag-bearer for Toronto artists is pop singer and songwriter Shawn Mendes, whose second album Illuminate has gone to number one in no fewer than 65 countries.

The majority of Illuminate was produced by Briton Jake Gosling, known for his work with Ed Sheeran, with more songwriting and occasional production from Americans Scott Harris and Teddy Geiger, amongst others. There also was top-level support from Toronto mixer and engineer George Seara, who’s been part of Shawn Mendes’ success story from the start. Seara recorded most of Mendes’ debut EP and his first album Handwritten; he also mixed several of the songs on the latter album, and mixed six of Illuminate’s 12 songs, including the lead single, ‘Treat You Better’, on the SSL console at Toronto’s Noble Street Studios.Following the likes of the Weeknd and Drake, Shawn Mendes has become the latest Toronto native to enjoy global success.

A Shared Aesthetic

The songs Seara mixed are all in the first half of the album, along with ‘Mercy’, which was mixed by the legendary Tony Maserati; the remainder of the songs were mixed by Jake Gosling and his engineer Jackson Dimiglio-Wood. Maserati’s mix has a certain pop gloss, the Gosling/Dimiglo-Wood mixes are a little rawer and more acoustic-sounding and Seara’s mixes suggest depth and sophistication as well as his own version of pop gloss.

“I’ve often heard from the label, and Shawn, that they really love the way I treat the vocals,” explains Seara. “I tend to get what it is that Shawn is looking for. There’s a shared aesthetic. So in general, I’m given some small notes beforehand and any mix revisions go smoothly. Prior to album mixing I also spoke with Jake [Gosling], and the main thing he said he wanted was to keep things ‘warm and dynamic’. It was great to hear that as it’s often my preference. It’s not uncommon for some contemporary pop projects to be really aggressive and loud, and that’s OK for some things, but it’s refreshing to keep things warm and dynamic, and this worked really well for Shawn’s record.

“Songs like ‘Three Empty Words’, ‘Don’t Be A Fool’, and even the album opener, ‘Ruin’, were recorded in a very acoustic way, with everybody set up in a room. Shawn was initially singing and playing guitar in the same room where the drummer, bassist and other musicians had set up. I think this approach helped in maintaining an organic feel for many of the songs. In today’s world, pop songs are generally constructed around programmed parts, and all instrumentation is overdubbed. Many of the songs started from an idea by Shawn — he’s really worked hard on his singer-songwriter skills — which were then developed by him and his team. I think that vibe carries through on the record, because that is how it was captured, and I wanted to maintain that by mixing in analogue.

“‘Treat You Better’ is probably one of the more pop tunes on the record. It was co-written by Shawn, Teddy Geiger and Scott Harris, and has many acoustic instruments, like the other songs, but in this case they are also supplemented by quite a bit of electronic stuff, to give it more of a contemporary feel. By putting that through a desk, and also by using a bit of outboard — I’d say 75 percent of the processing and effects I used in this mix were out of the box — I found that I could give things some depth, while maintaining a tight bottom end and silky presence in the top end.”

‘Treat You Better’ was mixed on the SSL E-series console at Toronto’s Noble Street Studios.

Smoothing The Rough

George Seara continues by describing his mix process for Illuminate in general and for ‘Treat You Better’ in particular. “I did all the mixes unattended. I’d start on a song at 10am, and generally I’d print my mix by 5 or 6 pm, and I’d then send it to the team. They’d live with it for a day or more, and would send me their notes. In some cases I’d receive the notes a few days later, so I would move to other things, including other songs for Shawn. I used desk automation and Total Recall, and also have much of my outboard routed permanently on the same channels on my desk, which makes recalling very easy.

“Some of the songs were recorded in Logic and some in Pro Tools, so in the former case I’d receive WAV files, consolidated and exported from Bar 1, Beat 1, which my assistant would then import into new Pro Tools sessions. In the latter case they sent me their Pro Tools sessions. I always ask for complete Pro Tools sessions, if and where possible, including the latest rough mix, any automation and plug-ins used, everything. People invest a lot of creative energy into developing these songs in the studio, and often they have been living with rough mixes for some time. As a starting place, it makes sense for me to have a listen to their direction and quite often the challenge is to retain that and at the same time refine and elevate the mix considerably. It’s not uncommon for rough mixes to come in already in a great place.

“It’s important to note that there are myriad possibilities with regards to the way that things are recorded and treated today. For instance, sometimes a guitar sound is developed by plugging a guitar directly to a UA Apollo, and then plug-ins are used to adjust the sound to taste. While I appreciate having a clean DI sound to work with, I also need all of the plug-ins, so that I have them as an optional starting place. Once again, the idea is to be mindful and respectful of the rough mix, and then ‘raise the bar’ considerably.”

Vocal Focus

Another characteristic of the modern, DAW-based way of working is that sessions often end up having tons of tracks, and the Pro Tools mix session for ‘Treat You Better’ is a typical example. It has 102 tracks, 89 of them stereo, though about 20 of these are stereo aux and VCA tracks added by Seara. This meant that in order to mix the song on the SSL, he was faced with the task of condensing 160-odd audio streams to the 44 channels that he had available on the desk.

“For this project, I had everything returning onto the desk in groups: stereo kicks, stereo toms, stereo claps, stereo cymbals and so on. I generally want to get everything on the desk as quickly as possible, so that I can play things back and hear the song as a whole, and can start pushing the faders up and down. I like to try and keep the faders in a nice working range, which means that I may trim things back a little bit in Pro Tools, or possibly use the analogue trims at the top of the desk. Those are my first steps, and I then get a quick, nice monitor balance. At this point I am not really soloing anything, but just getting a feel for things as a whole. Only after that will I solo tracks and channels, in Pro Tools and on the desk, to check out what I have, and what the plug-ins are doing and what’s needed. Sometimes I’ll decide that the plug-ins on a track are not needed, and I just use the desk and outboard.”

Many mixers work from the ground up, starting their mix with the drums before adding the bass and other instruments, and the vocals only after that. Seara, however, uses the reverse approach: “This may sound a bit backwards, but I generally start my mix with immediate attention to the vocals and whatever the key musical elements are in the song. Obviously the beat is a key component, but in the case of ‘Treat You Better’, for example, I began with the lead vocal and key rhythm guitars, and then brought in percussion and bass. I want the vocal to be warm and present, front and centre, bringing key instrumentation in with it, focusing on these elements first and foremost, developing a sound that feels really good for that, and then start to bring in other things around it, including additional percussion, guitars, synths and so on.”

Moving Quick

This composite screen capture shows the entire Pro Tools session for ‘Treat You Better’.As mentioned above, the Pro Tools session of ‘Treat You Better’ contains just over 100 tracks. At the top of the session is Seara’s stereo mix print, and immediately below that are 13 aux effect tracks, followed by 34 drums and percussion tracks, which include nine tracks of claps and four of finger snaps. Further down are five bass tracks, a whopping 28 guitar tracks, three synth tracks and a strings track. The 17 vocal tracks are split between five lead vocal tracks, three ad lib tracks, and eight backing vocal tracks, and a ‘Vocals ALL’ VCA fader.

“I use quite a few VCA tracks when I am mixing, for both organisation and control, in Pro Tools,” says Seara. “They are coloured red in my sessions, with main VCAs in this project being Drums, Bass, Guitars, Synths, Vocal Leads, Ad Libs, Backgrounds 1 and Backgrounds 2. Similarly, much of my instrumentation on the SSL desk is assigned to its centre section VCAs, ie. drums, bass, guitars, synths, backgrounds and so on. Once I have a nice balance on the desk and the mix is already feeling pretty good, I’ll begin with some SSL desk automaton where I’ll do some basic rides and mutes on the desk. Next I’ll begin to refine some moves in Pro Tools as well, and many of my moves in Pro Tools are done using these VCAs.

“I have an Avid Artist Mix set up just next the my computer keyboard, so I can quickly toggle Pro Tools tracks into Write/Touch mode and perform more automation moves. For fine detail stuff I do use my trackball quite a bit. The aux tracks at the top are set up with some default reverbs or delays of mine, just in case I need them. They include four delays from SoundToys Echo Boy, and four reverbs from UAD, the EMT 140, AKG BX20, and two Lexicon 224 reverbs. There were a few existing effect aux tracks just below these already in the session and I kept them available, so that I could audition them as well.

“The hardware TC Electronic System 6000 is my main reverb and effects unit and it has four engines [ie. four individual stereo effects units], three of which I have returning onto the SSL. I have Engine 1 set up with a Chamber reverb, Engine 2 with a Stage+Hall, and Engine 3 with a modulation effect, which is an AMS-type preset called ‘Pitched Stereoizer’ where the left channel gets pitched up slightly and the right channel gets pitched down slightly, and there are different minor delays in each channel, just to give things a little bit of modulation width. I use that sparingly to ‘widen’ or ’stereoise’ some things, including bass and lead vocals. I also always have a few go-to outboard pieces, like Empirical Labs Distressors, API 2500, Elysia Mpressor, ADR F760X ‘Compex’ and Pendulum OCL2 on a few bus assignments and inserts, so that I can use them at any time. I have the Pro Tools aux effects returning on the desk on a couple of faders as well.”Much of George Seara’s mixing work on ‘Treat You Better’ was carried out using outboard, including (from top): Lynx Hilo, Avalon AD2055 EQ, Empirical Labs Distressor and Retro 176 compressors (x2 each), Neve channel strips and UA 1176 compressors (x3).

• Vocals: SSL desk EQ & compression; Avalon 2055; Empirical Labs Distressor; Retro Instruments 176; TC Electronic System 6000; Waves Renaissance De-esser, SSL Channel, CLA Vocals & Puigtec EQP1A; McDSP FilterBank & AE400; UA Blue Stripe 1176A, Precision De-esser, AKG BX20, Manley Massive Passive & Fairchild 670; FabFilter Pro-Q2; Pendulum OCL2.

“The lead vocal for this session came as a stereo track, but it doesn’t necessarily mean it actually is stereo, it’s just that when people work in different systems, including Logic, and bounce things back and forth to different workstations, sometimes things end up as stereo tracks. I can either split them to dual mono, in which case I only use the left channel, or I leave them as is. In this case I did the latter. So on the desk I ended up with the lead vocals coming up on channels 17 and 18. In the session I split the lead vocal track over three tracks, because there were some words and phrases I wanted to treat differently.

“The hardware on the lead vocals includes SSL desk EQ, with a low cut at 70Hz to get rid of any rumble, and mild boosts at 150Hz and above 15kHz, to add some presence. I also used desk compression, set to 2:1, and my lead vocals had an Avalon 2055 stereo EQ going into a pair of matched stereo Distressors going into my matched pair of Retro 176 compressors. My Retro Instruments 176 compressors are matched-stereo and linked, and set to a ratio of 4:1 with medium attack and fast release, no inter-stage [transformer] and side-chain filtered at around 70Hz. I absolutely love these compressors on vocals. My Avalon 2055 EQ also had a mild presence boost above 15kHz. In addition I used some System 6000 Stage+Hall reverb and a bit of TC System 6000 ‘AMS vibe’ for ‘widening’ the vocals.

More of the outboard used by George Seara in his mixes for Shawn Mendes (from top): ADR F760-RS stereo ‘Compex’ and API 2500 compressors, Empirical Labs Fatso Jr tape emulator, Avalon AD2044, Teletronix LA-2A, UA 1176 and Empirical Labs Distressor (x2) compressors, Chandler Germanium Compressor (x2), Crane Song STC8 compressor/limiter and ADR F760X compressor (x2).“There are quite a few plug-ins on the lead vocal tracks, but I used each of them very sparingly. The signal chain on the inserts of the main lead vocal track is the Waves Renaissance De-esser, Waves SSL Channel, McDSP FilterBank, UAD Blue Stripe 1176A, FabFilter Pro-Q2 and finally UAD Precision De-esser. But again, it’s almost like I’m feathering things. I’d say that the most important ones were the 1176, Q2 and Precision. I really like the latter. It has a little bit of a sound, so sometimes it improves something just by sending audio through it. I also used a McDSP AE400 Active EQ on the vocal and it’s on an aux just above lead vocals. I’ve been working with that recently, and it works incredibly well as a dynamic EQ. Basically it allows me to add a bit of presence, and control any brashness. It also works wonderfully as a de-esser. Plus the lead vocal has a touch of the UAD AKG BX20 which is an incredible spring reverb emulation.

“I treated the ad libs also with desk EQ, cutting at 70Hz and boosting slightly above 15kHz, and desk compression at 3:1 with fast release and linked, and also sent the ad lib vocals to a Pendulum OCL2 compressor. Plus I used the same two System 6000 reverbs as I used on the lead vocals. The SSL channels here are ‘floated’, meaning that they are not directed to the mix bus, but rather to my Pendulum OCL2 compressor, and from there assigned to the mix bus. This allows me to send various channels to one stereo unit, in this case being all backing vocals and ad-libs. I again used several plug-ins here, though very sparingly. The most important ones are the Waves CLA Vocals, Ren Compressor, Waves SSL Channel, FabFilter Pro-Q2 and UAD Blue Stripe 1176A. I treated all backing vocals outside of the box in a similar way as the ad libs. The only difference was in the plug-ins, with the BG1 vocals having the McDSP AE400, Puigtec EQP1A, UAD Manley Massive Passive, Filterbank, UAD Fairchild 670, and Pro-Q2, and the BG2 the Filterbank and UAD Fairchild 670.”McDSP’s AE400 ‘active equaliser’ was used to add presence and control harshness in the vocal.

• Guitars: SSL desk EQ; Avalon 2044; Manley Massive Passive; TC System 6000; SoundToys Echo Boy; Waves Renaissance Axx.

“The main thing I can tell you about the guitars is that they have desk EQ, and outboard includes the Manley Massive Passive EQ, with a small boost at 200Hz, going into an Avalon 2044 compressor, and also using some reverb from the System 6000 Chamber. The Avalon 2044 comp is an opto comp and I really like the sound of it on guitars. It takes the ‘edge’ off digital recordings just a bit, and I used it gently with a 4:1 ratio, slow attack, fast release, couple of dB of compression. The main palm-muted guitar in the beginning also has a delay from SoundToys Echo Boy on it. I used an Echoplex setting with eighth-note delay, and tweaked the time of it slightly to ‘sit it back’ in the mix. I also cut some bottom end and high end using the low cut and high cut directly in the plug-in. I’m a big fan of the SoundToys Echo Boy plug-in and use it often, generally with some roll-off in the high end and low end. The saturation knob is also a useful feature and for this particular delay I set it around three o’clock (approx 19dB). I also used a bit of the Renaissance Axx on some of the chorus guitars to help ‘lift’ things a bit.”The upper rack in this photo contains the two Focusrite ISA 110 equalisers used on the bass, and an ISA 130 channel strip. The centre rack is home to Seara’s Manley Massive Passive equaliser and Chandler LTD2 compressors, while the lower rack houses (from top) Pultec MEQ5 and EQP1A (x2) equalisers, SSL FXG384 bus compressor, AMS RMX16 reverb, Airfield Liminator compressor and Drawmer MX50 de-esser.

• Bass: SSL desk EQ; Focusrite ISA 110; Airfield Liminator 2; TC Electronic System 6000; Waves Renaissance Bass & CLA Bass; McDSP FilterBank.

“Like everything else, I began with the SSL EQ on the bass, boosting 3kHz and filtering some low end below 27Hz. On the inserts I had a pair of vintage Focusrite Blue ISA 110 EQs, and then an Airfield Liminator 2 compressor, plus a bit of TC System 6000 for AMS-style widening. The Airfield is great for ‘invisible’ compression and I’m compressing about 2-4 dB. It has a mode that emulates a little bit of a transformer-type sound, and that is what I set it to. It just holds the bottom end together a bit more and helps make things sound a little more cohesive. I also love the Focusrite Blue EQs. They have a Neve sound to them, and I often boost the bass again around 3kHz with it, to help it cut through a little bit more.

“Plug-ins on the bass were the Waves Renaissance Bass and Filterbank and a little bit of the Waves CLA Bass. The CLA Bass was already in this session, and I kind of liked what was happening with it so I opted to keep it in. Basically it was adding a bit more bass and sub, and I cranked up the distortion parameter, set to ‘Growl’, a bit more. By the way, just above the drums is an ‘SC Key’ track which is used as a side-chain trigger. The idea was for it to trigger some instrumentation such as bass, some keyboards, and even a few background vocals to give things a little bit of a rhythmic push-pull ‘pumping’ effect. You hear this in EDM music quite often, though it’s used here much more subtly. This was something that was already set up in the session, and in the end I think I may have used it very sparingly in the mix.”

• Drums: SSL desk EQ; TC System 6000; Empirical Labs Distressor; API 2500; Elysia Mpressor; ADR F760X; Waves SSL Channel; SoundToys Decapitator; McDSP Filterbank; FabFilter Pro-Q2; UAD Lexicon 224 & AKG BX20.

“The drums and percussion returned on the desk in groups: kicks, toms, snares/claps/snaps, and cymbals and miscellaneous percussion. Sometimes I’ll move things around a bit, but generally speaking, things are broken out like this. Small faders on the desk were used to ‘feed’ various parallel compressors to taste. I started again with hardware SSL desk EQ, cutting bottom end below 27Hz on all drum channels, giving the kicks a small bump at 50Hz, and subtle boost around 3kHz and 8kHz respectively. I added a touch of sheen above 12kHz on the claps and snaps. I also used a bit of TC System 6000 reverb chamber on claps, snaps and cymbals.

“Using small faders I fed some drums and percussion to Distressors, my API 2500 and my Elysia Mpressor. My vintage ADR F760X ‘Compex’ gets used quite often on overheads, cymbals and room mics, and I had it on an insert with the cymbals and miscellaneous percussion. Everything starts on the SSL desk, and the Distressors and API 2500 are used to help bring the bottom end together and get things ‘hitting hard’ in a musical way, though I used the compressors in parallel so that the drums never lose their initial impact and explosiveness.

To help one of the kick drum tracks cut through, SoundToys’ Decapitator was used to apply distortion.“On the plug-in front, I used the Waves SSL Channel, SoundToys Decapitator, McDSP Filterbank, FabFilter Pro-Q2, UAD Lex 224 and AKG BX20. I love the Waves SSL Channel and like it for quick EQ and adjusting overall level via its trim. The SoundToys Decapitator is a great tool for more impact and aggressiveness. I used it on the ‘Kick2’ track in the break just before the final chorus to help create an even more aggressive and explosive sound. The Drive control is up pretty high and I’ve pulled back the overall output level in order to ‘level match’ things. I use the McDSP Filterbank often to remove some extra-low bottom end that is not needed and it can help in tightening things up and ‘defining’ the overall bottom end. On drums I typically cut between 30-45 Hz. On the snaps and claps I again used the Waves SSL channel and Decapitator, and some desk stuff.”

• Stereo mix: Millennia NSEQ2; RND MDB; SSL desk compressor.

“On my master bus on the desk I had my Millennia NSEQ2 with Fred Forssell modification, going into a Rupert Neve Designs Master Buss Processor, followed by the SSL desk compressor. My Millennia NSEQ2 has been modded so it no longer has a tube stage as an option. It’s a great-sounding EQ. Its JFET stage sounds incredible, with silky highs and tight lows. I had it set to a very small boost at 50Hz and a very gentle presence boost above 15kHz. The RND Master Buss Processor has become another favourite of mine and I’m using it in ‘feedback/RMS’ mode with slow attack and fast release, 2:1 ratio, no limiting. I’m also adding just a touch of ‘red’ silk. Plus, I’m using a bit of SSL desk compressor with slow attack and auto release, 4:1 ratio.

“The stereo mix gets printed directly back into the session via a Burl B2 Bomber A-D converter and I print quite a few stems as well. When I send my mixes out for feedback to clients I do bump the level up so that it’s much easier to reference alongside other commercially available material. I’ll sometimes shoot out a few different options for bumping level up and see what works best. It’s not always the same plug-in and same setting, and some of my favourite plug-ins for bumping level up include iZotope Ozone, McDSP ML4000, Slate Digital FG-X and Voxengo Elephant. Also, I have an entirely independent computer running Magix Sequoia or another instance of Pro Tools that I use to print final mixes simultaneously via my JCF Audio Latte A-D converters and a Lynx Hilo, at 96kHz or 192kHz. These mixes are of the very best quality and we master these, if and where possible.

“I have to admit to being a bit of a gear aficionado, though I really challenge myself to only add gear to my racks that makes things better. Funny story, I’m a big fan of the legendary Al Schmitt and his fine engineering work, and I learned about the JCF Audio Latte converter by way of an album project that we both happened to work on, though in different cities. When I heard about the converter, I tried it out and fell in love with it. There was a time when I used to print mixes to half-inch analogue tape. This is simply not practical any more today, so I do it very rarely these days. That being said, this converter sounds so great to me, that it’s another one of the reasons why.

“Does it make a difference to go for 96 or 192 kHz? At the end of the day, trust your ears. If I have mixed via analogue in some capacity, on a desk or using some kind of summing box, and I print both a 44.1kHz mix and a 96kHz mix, it really depends on the calibre of converters that are used, but I’m very confident that you would hear the difference very clearly on a decent playback system. But leaving all of the tech talk aside, all of us need to make our decisions based on what is the most efficient, what is the most practical, and on what serves the song best.”