Engineer and producer Dani Castelar accompanied Paolo Nutini through three years and 11 studios to make his acclaimed new album Caustic Love.
Scottish singer-songwriter Paolo Nutini has explored an incredibly diverse range of music styles during his career, including ska, reggae, swing jazz, doo-wop, gospel, '60s psychedelic, calypso, soul, skiffle, Frank Sinatra-style crooning, blues, New Orleans, funk and rock. The results have brought him both critical and commercial success: debut These Streets, released in 2006 when he was still only 19, and the 2009 follow-up Sunny Side Up have each sold more than two million copies worldwide.
Released a whole five years after his second album, Caustic Love has been a UK and Irish number one, and has charted even higher than his previous albums in many other places around the globe, from Australia to Austria. The US has yet to succumb to the charms of Nutini's voice and music, but this may just be a matter of time, especially as Caustic Love is heavily influenced by '60s and '70s American black music. As the album's title suggests, it's a much darker record than the optimistic Sunny Side Up, comprising mainly moody, atmospheric ballads and deep, mid-tempo grooves. The low-down, gospel-influenced swamp-rock-funk of the first single, 'Scream (Funk My Life Up)' is the most up-tempo track on the whole album.
Caustic Love took three years to make — a long time by modern standards. This reflected Nutini's search for a new direction, in terms of his songs and their production. Caustic Love has been welcomed as his coming-of-age album, marking his graduation from adolescence to manhood and to artistic independence. It is telling in this context that big-name producers did not play a dominant role on his new album. While his debut saw the involvement of studio veterans Ken Nelson and Tom Elmhirst, and his second album was co-produced by Nutini with Ethan Johns, Caustic Love is for the most part a collaboration with the relatively unknown Dani Castelar, who gave Nutini the space to freely develop his own identity and direction.
Dani Castelar is originally from Valencia on the east coast of Spain. A year ago he returned to the relative quiet of his early-life home town after a tumultuous decade in Ireland and the UK. He played guitar as a teenager and used, he says, "early Logic and Reason to record friends and so on. I also made electro-acoustic music with beats and guitars.” Castelar went on to study political science at the University of Granada, but on realising that he wanted to make a living from music, he moved to Dublin where he studied at the Sound Training Centre. After graduating in 2005, he became an assistant at Ireland's premier residential studio, Grouse Lodge, quickly graduating to house engineer. Then, in late 2006, he received the kind of call that up-and-coming engineers may occasionally dream of, but only when they're in a wildly optimistic mood or under the influence.
"The owner of Grouse Lodge, Paddy Dunning, called me and said, 'Dani, you need to come to the studio now, because a very well-known artist has just booked the studio.' When I asked him who, he said, 'Michael Jackson'. I was like, 'No way!' I went to the studio, and the next day a massive white van showed up, and Michael Jackson walked out, carrying a Casio keyboard. At three in the morning I received a phone call from Michael, saying that he wanted to record some vocals. He ended up staying at the studio for four months, and I was his engineer during that time. The first two months it was just him and me, recording vocals over some tracks he had brought with him on a DVD. He then called in some producers, like will.i.am and Rodney Jerkins, who came to co-write with him. Nothing of these sessions has been released so far.”
With the mother of all credits under his belt, Castelar went on to work with several other well-known names at Grouse Lodge, including REM, Snow Patrol, Bloc Party, Editors, and others, often working with producer Garret 'Jacknife' Lee. "In 2007 Paolo Nutini showed up,” Castelar recalls, "and I engineered his Sunny Side Up album. He then asked me to become his regular engineer, so I went independent and moved to London. I have since worked regularly with him, while also recording and mixing others. Working with Michael Jackson for four months was probably my breakthrough as an engineer, while co-producing Caustic Love with Paolo has been my big break as a producer.”
The making of Caustic Love involved recording sessions in no fewer than 11 studios in the UK, Ireland, the US and Spain. Amongst them were RAK Studios and Trevor Horn's SARM studios in London, The Toy Room in Brighton, Sunset Sound in LA, and Grouse Lodge in Ireland. The recording budget for Caustic Love must have been huge, by modern standards. Castelar notes: "The recordings were not all the time. We would work a month, or two, and have a break and then come back. But yes, renting big studios like that was very old-school. In fact, I have never worked with anyone who was doing things as old-school as Paolo. We did think about buying our own equipment and setting up our own place, but we wanted to do things in a practical way that would allow us to focus on the music.”
The only time Nutini and Castelar went for a DIY approach was right at the start of the project, at the end of 2011, when they set up a mobile studio in an abandoned police station in Nutini's home town of Glasgow. "It's a derelict building owned by a friend of Paolo's,” explained Castelar. "We spent three months there, writing and developing songs and recording drums, bass, guitars and keyboards. Paolo's band came in, but he also played guitars, piano and Wurlitzer. The building had a shooting range in the basement where we recorded vocals. The gear we had was a Mac Pro with Avid HD interface, four Neve 1073 preamps, two DAV Broadhurst Gardens No 8 preamps, four Universal Audio 1176 compressors, and mics like AKG D112s and S414s, Shure SM57s and SM7s, Sennheiser MD421s, Neumann U87s and a few Samson DIs. The usual.”
Some of the material from the early sessions made it to the final mixes, and the album's closer 'Someone Like You' was recorded at the old police station in its entirety. In the early stages of the making of the album, Nutini was still looking around for big-name co-writers and producers, and as a result, the next stage involved a month at RAK studios in London with Leo Abrahams (Brian Eno, Imogen Heap, Jarvis Cocker, Paul Simon) producing and the great Tchad Blake and Castelar engineering. Material from these sessions surfaced on several of the album's 13 tracks, with three songs co-written by Abrahams and Nutini and co-produced by the two with Castelar: 'One Day', 'Superfly (Interlude)' and 'Fashion'. The last of these also featured Dave Sardy (Nine Inch Nails, the Stones, LCD Soundsystem, Red Hot Chili Peppers) as a fourth co-producer following the next batch of sessions, which took place at Sunset Sound with Sardy producing and Sardy's engineer Ryan Castle and Castelar engineering. Elements of the first single, 'Scream (Funk My Life Up)', came from these sessions, with Nutini and Castelar credited as producers and Sardy with "additional production”.
Castelar elaborates on how the material evolved through these sessions. "Both Dave and Leo are amazing producers, and they both did some pretty cool stuff but somehow it did not quite work out, and in the end Paolo and I returned to the approach that we liked best, which was just the two of us working together in the studio. So we ended up producing the record ourselves, using some of the things that we had done with Dave and with Leo, and putting them through our own filter. After the LA sessions we took a long break, and went back to writing mode. It was not like we were not happy with what we had, it was more that Paolo wanted to dig inside a little bit more and write more material. We did some of this at Grouse Lodge, where we spent a few weeks writing, and where Paolo wrote parts of 'Better Man' and 'Looking For Something', and where we recorded the song 'Cherry Blossom' live in the studio with Paolo's band. That studio has an amazing atmosphere when you're looking for inspiration.
"Following the Ireland sessions we went to SARM Studios in the West End. We love that place and made it our home base for a year, off and on, during which time we also went to other studios, like the Strong Room in London where we did some vocals, RAK, where we recorded 'Better Man' and 'Looking For Something' in two days, again mostly live, Studio Barxeta near Valencia, where we spent a couple of weeks doing vocals and editing, plus the Toy Rooms, where did more vocals. There also was a string session at Studio 45 in London with four string players who were overdubbed three times to get the effect of a string orchestra, but I wasn't present for that.
"In SARM we recorded mainly overdubs, like brass, guitars, vocals, keyboards and so on, and I also mixed four of the album's tracks there: 'Bus Talk (Interlude)', 'Superfly (Interlude)', 'Diana' and 'Someone Like You'. The rest of the songs were mixed by Craig Silvey, who added a certain magic to my rough mixes. Some of his mixes are my monitor mixes pumped up! His old-school approach with his vintage Neve and outboard was a perfect fit with the music and the sound of the record. Having said that, the '60s and '70s influences in the sound were not 100 percent deliberate. We played a lot of it by ear and didn't go in with a preconceived idea of what we wanted the whole album to sound like. The music speaks for itself, and all I had to do is guide it. The way I recorded and processed some of the things on this album had a strong vintage flavour, not because of some conscious decision, but because I felt that the material deserved it.
"In general I quite enjoy the old-school approach, and I used the desks in the different studios where we worked extensively, particularly the '70s API at RAK and the J-series SSL at SARM. I worked in Pro Tools at 24/48, don't ask me why, it was just an almost random choice at the beginning of it all and it stayed. I also mixed through a board, using a blend of outboard and plug-ins. We did do a few things to add an old-school flavour, like I recorded 'Numpty' to Pro Tools, transferred the recordings to tape and back to Pro Tools again. With a few other songs, like 'Better Man' and 'Looking For Something', I just dumped the drums on tape, and then brought them back into Pro Tools to add some tape saturation. I used a Studer A800 at 15ips, and I can hear the difference.
"We did a similar thing with the strings and brass, but instead of transferring them to tape, Paolo had the idea of having them cut to vinyl. This was done at Abbey Road Mastering, and we then recorded them back into Pro Tools. It didn't do a lot, other than add some crackle that you can also get from some plug-ins [laughs loudly]. But it feels really good when you do it. If you enjoy that process, it somehow makes you happier and gives you a satisfaction that you don't get when you use a plug-in. But the end product is pretty much the same. I think it is the skills that matter, not the tools. It's the driver, not the car. You can do a record completely in digital nowadays and make it sound as if it was done on analogue. To achieve this, there are multiple plug-ins, one in-the-box option being the Slate Digital plug-ins which are incredible. The Virtual Console Collection does exactly what it is supposed to do. It's not that you put them on an insert and you go 'wow'. Instead it's very subtle, and it adds a certain character.”
'Diana' is a slow-burning ballad, with a chorus sung in falsetto voice, a repeat echo (from a Boss RE20 pedal) guitar part played by Dave Nelson, who also played bass, a lone trumpet by Ben Edwards, Nutini on Wurlitzer, Gav Fitzjohn on piano and Seb Rochford on drums. There's a lot of space in the production and arrangement, reminiscent of the cool jazz-pop Sade records of the '80s. The song and production are a masterclass in achieving maximum mood impact with a minimum of ingredients, but according to Castelar, it almost wasn't developed.
"The song was written during the second set of sessions at RAK, and we recorded the first takes there one morning when we all had a hangover, laying down a Wurlitzer, a guitar, the bass and a guide vocal. The Wurly, first guitar and bass still remain on the final album. It was very ad hoc, and we never considered developing the idea, but everyone who heard the song loved it, so eventually we did continue to work on it. This mostly happened at SARM, where we overdubbed vocals and keyboards, another guitar, drums, and some drum programming, and Paolo overdubbed yet more vocals at the Toy Room.
"My approach when recording drums is pretty standard, but for this track I used a very minimal approach. This was influenced by having watched Tchad Blake work during the earlier sessions at RAK. I tried to learn as much from him as I could, because he is a master. I kind of kept his approach for recording the drums on 'Diana', which meant that I had an AKG D112 and Neumann U47 on the kick, a Neumann U87 on the snare, an RCA 44 as a mono overhead, and that was it. On other tracks, like 'Better Man' and 'Iron Sky' I had many more drum microphones. There's also a rimshot sample on 'Diana' of a sound that we had recorded ourselves, with an SM57.
"The bass on 'Diana' was recorded DI and with a Neumann U47 FET on the amp, and the guitar cabinet with a Shure SM57 and a Neumann U87. Paolo's Wurlitzer went through a Roland Cube and was recorded with an Shure SM7. The way I record is to just grab what's around in the studio. If the inspiration is there, you record with whatever you can. Don't spend a fucking hour making it sound good! All mics went through the API desk at RAK. I used a Shure SM7 on Paolo's voice, and a Neve 1073 going into an old Urei 1176 'blackface' [compressor]. Paolo usually sits beside me when he sings, with a pop filter in front of the mic. No fancy mic booths. This allows us to discuss things easily, rather having to use the talkback mic.”
"Although the plan was for Craig [Silvey] to mix all the tracks, I did sit down to mix a few tracks properly, laying them out over a board and trying to take them beyond my monitor mixes. 'Diana' was one of them, and the mix I did had something sexy that we liked, so we decided to keep it. I approach every mix I do differently, to make sure there's something new and fun to keep me interested. In general I do my basic mix in the box, and I then spread things out on the board, for summing and to add some outboard processing. With 'Diana' I started the mix in my hotel room in London, using a laptop and headphones. I started from my monitor mix, because that had a vibe, and I tried to see where I could take it next. I bounced some of the track and added new plug-ins for EQ, compression, reverb and so on. The next morning I went to SARM and laid it out over the board, first with all the faders to zero, and then I started doing my moves on the SSL from there. After that I would have gone back to Pro Tools to do some very detailed volume automation moves.
"When I begin a mix I tend to listen to everything at the same time, and I then take it back and work on the individual elements, usually starting with either the drums or the vocals. On this track I probably began with the vocals and not the drums, because the feel and atmosphere of the vocals were the most important elements. I EQ'd and treated the vocals and I then started adding things around the vocals. With Paolo's music his vocal usually is the centre of everything. I really like his vocals to be up front and in command. So I worked on the vocals until they felt right, and after that I just went with that.
"You can say that doing a final mix is like fine-tuning the rough mix. With 'Diana' we liked the rough very much, so it was just about making it bigger and making it sound more like a record. Having said that, I'm not too big a fan of the idea of 'making something sound like a record', because with some professional mixes, you don't hear the musicians playing any more, you just hear a very professionally presented palette of sounds, very well put together, but you can't picture anyone actually playing these sounds any more. I can see the point, but I'm not a big fan of those mixes. The monitor mixes had an element of aliveness, with you being able to hear real musicians playing, and I really like to maintain that. When I mixed 'Diana' there was of course an element of making what we had done a bit more accessible to the general public, but mostly we were concerned with keeping the atmosphere dark and intimate, as opposed to it being a bright pop mix.”
- Drums: desk EQ and compression, Bomb Factory BF76, Waves API 550, CLA Drums, Q4, API 2500, Kramer PIE, Puigchild 670 & SSL compressor, SoundToys Decapitator, Avid EQ III, Audio Ease Altiverb, Avid D-Verb, Chiswick Reach compressor.
Castelar: "The verses in 'Diana' had a very minimal approach: just a kick, a rimshot and some hats, and nothing else. I recorded these elements separately, and placed them in the verses, programming them so they were in the right place. I'm old-school. I record my own samples and I then process them, and bounce them and manually place them in the Edit window. Track number 3 is the kick drum that sounds in the entire track, in stereo, because I bounced it down from the two kick mics, the D112 and the U47, and I would, in the process, also have used some plug-ins. There's a BF76 compressor and the API 550 EQ, just for some light compression and adding some bottom end. When bouncing I might also have used some kind of reverb to add some width. Track 4 is a kick sample, which has the Waves CLA drums and Q4 plug-ins, probably for some more bottom end as well, plus I will have cut some frequencies with the Q4. Track 5 is also a sample, but I didn't use it. Track 6 is the rimshot sample, which has the Decapitator, pushing it hard to add some more harmonics, and the Avid EQ, to cut out all the low end. It also has the API 2500 compressor and a spring reverb from the Altiverb.
"Then, in the choruses, the verse drum elements are joined by the full drum kit, recorded with the four mics. I processed the drums in the choruses differently. I recorded the chorus kick mic through the old Shure Level-Loc [compressor]. Track 8 is a drum sample that only happens in the choruses, probably to beef up the snare, on which I had the Waves Kramer compressor. Track 9 is the hi-hat in the second verse; it's a loop I programmed from one of my own samples, and it has the Waves JJP 670 Fairchild compressor and the API 550 EQ. I'm not compressing very hard, just adding some grit. Track 10 is a ride cymbal and tracks 11-14 are the acoustic drums recorded with the D112, U47, U87 and RCA44. I didn't treat them in the box, and probably had SSL desk EQ and compression on them. Tracks 15 and 16 are drum submix tracks, on which I again have the JJP Fairchild and the SSL bus compressor plug-in. Track 18 is a drum aux with the D-Verb on it, to which I sent the verse rimshot and the chorus snare. I also had a Chiswick Reach compressor on the drum group on the SSL desk.”
- Keyboards: Avid AIR Stereo Width, Waves DeNoiser, NLS Channel, Aphex Aural Exciter & MV2, SoundToys Echo Boy, Avid EQ III.
"Track 19 is a sample from a Yamaha DX7 or CS80, I can't remember, that has a stereo-widening plug-in. Tracks 20 and 21 are a real piano, which I recorded with Shure Green Bullet harmonica microphones. I compressed the mics really hard with an 1176, and reversed the piano in the box, so I didn't have to do anything else to the piano sound in the mix. Track 22 is a piano bounce. Tracks 23-26 are Wurlitzer tracks. There are three Wurlitzer tracks, two that play in the verses and one in the choruses. We loved the sound of the Wurlitzer going through the Roland Cube, but the resulting sound was very noisy, and while we liked that too, with three tracks of Wurlitzer the hiss became unbearable, so I took the hiss out on two of the parts and kept it on one. I just used the Waves DeNoiser plug-in, it was very easy.
"I used quite a few plug-ins on the Wurlitzer tracks, including the Waves NLS Channel, to give it some of the same harmonic distortion you get when you send something through a desk, the Echo Boy set to the 'Ambient Echo' preset, to add a bit of a dreamy feel, and the Waves Aphex Aural Exciter for some more top end, some additional air. I also have a high-pass filter from the EQ III, and use the Waves MV2 multi-band compressor to compress the high and low end. I love that plug-in because it has just a few knobs. I love gear with few knobs. It sounds better somehow. You need to spend a lot of time with gear with many knobs, and when you are recording or mixing you often don't have that kind of time.”
- Vocals: Bomb Factory BF76, Waves L1, API 550, Renaissance De-esser, Renaissance Vox & JJP Vocals, Avid EQ III, D-Verb & AIR Stereo Width, GML EQ, Tube-Tech CL1B.
"Tracks 27-28 are the verse lead vocals — Paolo sang them twice — and 29-30 and 32 the chorus lead vocals. I have the BF76 and L1 on the verse lead vocals because we recorded his vocals very softly, so he could sing very intimately. I did not want to process his vocals a lot. I wanted Paolo's voice to sound very natural, like somebody softly singing in your ear. So I added some mild compression and a bit of limiting, and some top end from the API 550 plug-in. The chorus vocals have the Waves Renaissance De-esser and the Avid EQ, which is adding some air at 20k. Paolo's vocals are so amazing, I barely have to touch them. I did send them out to the desk and EQ'd them with the GML EQ and the Tube-Tech CL1B compressor, but using these only very lightly. Track 35 is a vocoder vocal that only happens in the first verse, using an old Korg VC10. The backing vocals are on 36-39 and are bussed to tracks 40 and 42. I have the D-Verb on track 40, and on track 42 the EQ, the Waves Renaissance Vox for some compression, the JJP Vocals plug-in and the Air Stereo Width.”
- Other instruments: Avid D-Verb, Waves API 560, Renaissance Vox, Q4 & MaxxBass, Bomb Factory SansAmp PSA1, Camel Audio Camel Crusher, SoundToys Decapitator.
"Track 41 is the trumpet, on which I had just a bit of D-Verb. Track 43 is the bass, and I'm using the API 560 EQ, the Waves RVox, Q4 and Maxxbass. I'm compressing the hell out of the bass with the RVox. The two guitar tracks have the SansAmp, Camel Crusher, SoundToys Decapitator, and the D-Verb. The guitar was recorded rather cleanly, and I used the SansAmp to get more drive and crunch, and the same with the Decapitator. The Camel Crusher is set to 'More Beef' for… more beef! That plug-in is a freebie and it sounds amazing.
"I mixed back into the session, and had an SSL Quad compressor and another Chiswick Reach valve compressor on the stereo mix, just compressing things mildly. I don't like to over-compress mixes. Things have to be able to breathe!”
The Pro Tools project for 'Diana' contains 35-odd audio tracks, a relatively small number by modern standards. There are two click tracks at the top, followed by 12 drum tracks, seven keyboard tracks, 12 vocal tracks, a trumpet track, a bass track, and two guitar tracks. Right at the bottom is the stereo mix. As Castelar noted above, the low track count is in part due to him having bounced some of the recordings. In terms of organisation, it's notable that all tracks are numbered, but that no colour-coding is used. Castelar explains: "I would have had track colour-coding during recording and editing and mixing in the box, but I took it off when I laid the mix out over the SSL, because I wanted the Pro Tools session to be like a multitrack tape recorder. I didn't want the distraction of looking at a screen with all sorts of attention-grabbing colours.”