Inside Track: Lady Gaga 'Rain On Me'

Lady Gaga's latest album was two years in the making, and marked a return to her electronic roots.

Released on 29th May 2020, Lady Gaga's sixth studio album Chromatica has been overwhelmingly successful, going to number one in dozens of countries, including the UK and the US. The single 'Rain On Me', featuring Ariana Grande, also topped many charts.

Musically, the album is a throwback to '90s house music, set in a modern pop and EDM-inspired context. It sees Gaga reconnecting with the electronic dance and pop of her first three albums, The Fame (2008), Born This Way (2011) and Artpop (2012). Since then, she made an album with Tony Bennett, Cheek To Cheek (2014), a solo project inspired by Americana, Joanne (2016), and then there was of course the smash soundtrack for A Star Is Born (2018).

Chromatica is a big-budget album that was made over two years, with the involvement of dozens of musicians, producers, engineers, arrangers and programmers. Producer Bloodpop was Gaga's main co-writer and collaborator, while other writers and producers include Matthew Burns, the legendary Max Martin, Tchami, Rami Yacoub, Madeon and Skrillex.

Roll Call

Benjamin Rice has worked with Lady Gaga since her 2012 album Artpop.Photo: Buffy HubelbankTwo other essential names in the album credits are those of Benjamin Rice and Tom Norris. "I first worked with LG on Artpop," says Rice. "I started out as second engineer, and towards the end of the project I became first engineer, and from that point I have been her engineer, mixer, vocal producer and so on. I handle many different responsibilities, rather than these being split up between different people. When you become someone's trusted collaborator, you provide them with consistency, you understand what they are going for, and you become part of their creative direction. It's a service job, and it's an honour to do it.

"I was with her for the entire two years that it took to create the A Star Is Born soundtrack, and the same with the Chromatica album, from the first day of writing until the day we were approving masters. I'm also good friends with Bloodpop and we have worked together on some other projects as well. For the writing of Chromatica we'd go to LG's studio, which is Frank Zappa's former Utility Muffin Research Kitchen studio in the Hollywood Hills. [See the Craig Parker Adams interview elsewhere in this issue — Ed.] I try to keep up with everything: recording all ideas, recording vocals, starting to shape things, anything that she needed and the producers needed. I hopefully was the anchor of the situation.

"As the project progressed, we moved to EastWest Studios for a few weeks, then we spent a few weeks at Westlake Recording Studios, we were one day at Max Martin's MXM Studios, and we finished the album during a six-month spell at Henson Recording Studios. It took two years in total, though there were some breaks. Throughout, the core team for Chromatica consisted of Bloodpop, Burns, Rami, Tom Norris, who came in to mix during the last six months, and myself. We had more rotating producers who came in to collaborate, and I was in Studio B for vocal production.

"Every time there was a change to the instrumental productions, I'd try different vocal productions. There was a lot of trading files back and forth. It really was a team vibe. It felt like being in a band. I also recorded all the live instruments on the album: guitars, drums, percussion and so on, apart from the recordings of the three orchestral tracks. These were done at Henson, by composer Morgan Kibby, and she brought her own production team and players."

Meteoric Rise

Rice is originally from Texas, where he started playing drums, guitar and piano at age 12. By 13, he had fallen in love with the recording process. At 17 he got a job at SG Studios in Fort Worth, where he was mentored by producer Greg White, and he went on to play drums in the band Pilotdrift, touring the US for years. After a stint in the army, Rice continued his music career studying at Full Sail, and via Josh Gudwin (Justin Bieber's engineer, mixer and producer), Rice got a job at Record Plant in LA. It is here that he met Lady Gaga. Their collaboration eventually led to him earning an Album Production credit and Grammy Award for his work on the soundtrack for A Star Is Born and co-producing 'Shallow', Gaga's famous duet with Bradley Cooper and the most awarded song in music history.

Rice's wide-ranging experiences have led him to develop an approach to vocal production that's as unique as it is complex. His Pro Tools vocal recording template comprises 60 tracks or more, and he obsessively polishes hundreds of vocal takes and tracks to arrive at perfectly produced and mixed vocals.

"My philosophy with LG, and with this project in particular, is that I don't categorise vocals as rough writing vocals or final vocals. Every vocal that I get from her, whether from an early writing session, or on the last day of recording, has the potential to become the final take for that part. It really is an evolving process of chasing the target of what she wants her vocal to sound like, also depending on the production, of course.

"Whether we are starting a track from scratch, or continue working on a song that's already fairly well-established, I start with the same vocal mix Pro Tools template, which serves two purposes: it gives me a way to immediately get a good-sounding vocal with great effects, and that means that she can get really inspired in her headphones as she's singing. Secondly, it helps me with organisation and efficiency. So in effect I'm kind of vocal mixing on the fly."

Benjamin Rice: "I don't categorise vocals as rough writing vocals or final vocals. Every vocal that I get from her, whether from an early writing session, or on the last day of recording, has the potential to become the final take for that part.

Vocal Point

"I don't like the philosophy of recording vocals dry and boring, and then figuring out a sound later," continues Rice. "I like to figure out the sound as we are laying down the vocals. Why waste time? If we're writing, LG may be at the piano, which I'll record, and she'll maybe record a scratch vocal at the same time, so we can get the arrangement worked out. After that she'll do a piano take for real and we record vocals.

The input and output settings on Lady Gaga's vintage 1073 preamp."I have been using the same chain on LG for years, which starts with her personal, vintage Neumann U47, which I think is 60 years old now, and that has a great sound. We're used to it, and she loves singing into it. That goes into her vintage Neve 1073 mic pre, and then into a Tube-Tech CL-1B compressor, and straight into Pro Tools. It doesn't matter where we are, whether it's a studio, or at her house, or on a bus, I travel with this chain and record her with it. It makes it easier for me to punch in and out and combine takes from different days and places.

"With sessions for which the song is written and the artist is coming in to cut final lead vocals, I tend to have a more professional situation with the singer behind the glass in a booth. But with Gaga, if we're starting from scratch writing a song, I work differently, because we're recording as we're creating. It's common that we're both sitting in the control room next to each other, wearing headphones, or she's standing next to me, and that makes it easier to communicate as we try things. Or she may go into a booth. It depends.

"In my template I have lead tracks that go to a lead bus, and lead vocal double tracks that go to doubles buses and the same with backgrounds and ad libs. I have many internal sends and returns active and ready that I can change as I am working. I have plate, room and church reverbs, and eighth- and quarter-note delays that also have reverbs on them when I need them. There's also some dotted stuff, and chorusing effects and so on. I may not use them all, but they are available to me. I'm chasing what I think sounds interesting while she is performing. She's hearing these things, and will start to perform into these effects.

"I have a lead cut track, and eight lead cut tracks underneath that. She will typically sing a song from top to bottom, though she may stop to do just phrases, or verses or choruses. I organise all these takes by section, so when I am comping and need to reference something it's easier to find things, rather than having one track with 300 takes on it. My starting session may be 60, 70, 80 tracks of internal tracks, busing and organisation. To make sure the session remains manageable I work with only a two-track of the instrumental in my session.

"Every day that we cut a vocal, afterwards I go through all the takes, and make new comps and create a vocal production that could be a final version. I will spend hours polishing and refining ideas, find new ear candy and do other moves to enhance the aesthetic of the vocals. Once I feel I have the comps in a really good place, I listen to them on loop over and over again for anything that distracts me and can be better. I'll continue working on a session over weeks or months until for some reason I need to truncate it, and then I stem things out to something simpler, so I can add new things."

More To Do

Amazing as it may sound, at this point Rice is still far from done with his work on the vocals. "Once I have finished comping, I work on timing, expression, cleaning up doubles and harmonies, and I'll create a vocal arrangement, which can change greatly over time. When I feel I'm in a really good place with the comps and the vocal arrangement I export every vocal, topped and tailed to the same length, and bring these into a stand-alone Melodyne session, together with a two-track of the instrumental.

"For me, tuning is like photo-correcting. The photo is already beautiful, and you just touch up a few spots to make it even better. Tuning enhances an already beautiful and great performance. I use Melodyne not just for tuning, but also for improving the timing of the vocal, and I match the doubles and the harmonies to the timing of the lead. It is visually easier for me to do that in Melodyne. I'm doing this with dry vocals; there are no effects in my Melodyne sessions. I just focus on the a cappella vocals. I am also ducking esses and changing volumes across the comps.

"After finishing, I'll export all of those vocals and I bring them back in the song session, on a new playlist, on the same tracks on which I had the original comps. So now I'm playing the record with what I consider the final polished vocals, on their own playlist. If I want to reference a different take later, or if I want to reference the comps before I tuned them, they are all in the playlist underneath. After this point I will focus just on vocal mixing; how wet I want the vocal to be, what reverbs I like, and I'll do automation.

I use Melodyne not just for tuning, but also for improving the timing of the vocal, and I match the doubles and the harmonies to the timing of the lead. It is visually easier for me to do that in Melodyne.

"This leads me to a place where I'm close to the final presentation of the vocals. The whole process is in layers. Over the course of two years we may decide to change the production of a song, or she might come in and want to re-perform a song and see if she can approach it in a different way. When that's the case, I repeat the whole process again. But I will keep everything that I have done. If they change the production entirely, I may start a new session, but I'll still import all of my previous stuff, so I can have it as a reference, and I have the option of using old leads as new doubles to a new lead. It gives me the flexibility to keep building on what we have already done.

"Every song on the album is essentially a two-year session. This means that I end up with multiple sessions for each song. It'd be impossible to have one master vocal session that has every vocal take, every tempo change, every production change over two years. The session would just become too large. So I organise things by pre-production sessions, writing sessions, working sessions, final production sessions, mix-prep stem sessions, and deliverable sessions. It is a lot of organisation to keep up with things for two years!"

Co-mixing

Rice's vocal production and mixing process is clearly very involved. It's not surprising, therefore, that Lady Gaga's team decided to draw in another mixer to help out with the final mixes. Enter Tom Norris, a Grammy-nominated mixer and producer who has a big reputation in EDM and pop music, working extensively with the likes of Zedd and especially Skrillex.

Tom Norris (left) with Matthew Burns during the Chromatica sessions.

"A decision was made from the beginning that Tom and I would co-mix this album," recalls Rice. "Gaga is very happy with the vocal sound and mix I get for her, so why send the sessions out to someone who may put the vocals through more changes and unnecessary revisions? Bloodpop really wanted to hire Tom, because he is such an expert in dance music. He has a great understanding of the sonics of dance records, whether it's modern-sounding, like a Skrillex record, or more of a retro sound, like the Euro house stuff. So we decided that I would be responsible for mixing vocals, and Tom for mixing the music.

"I've been good friends with Bloodpop for several years," says Norris, "and he had been telling me for a while that he was working on the LG6 project with Gaga. I heard demos from time to time at his house. In January 2019 he asked me if I could try some production stuff on one of the demos. Fast-forward six months, he hit me up again, and said that they were beginning to finalise songs at Henson. The idea was to have Ben mix the vocals, and I would mix the instrumentals. I tried out one or two songs initially, and they liked what I did, so I ended up moving my studio setup there to do the full album. It ended up being from August 2019 to early March of this year.

"I was in the mix room at Henson, and Ben was in studio B, so we were right next door to each other and we could walk into each other's rooms to hear what the other was working on. It was a really fun, seamless process. Ben did his vocal mix with the two-track from whoever the track producer was, and I would receive a WIP a cappella bounce from him, and the stems of the instruments from the producer, and start on my mix. I'd then send Ben a two-track rough mix of the instrumental, and he'd load that into his session, and make adjustments, and the cycle would repeat until everyone signed off. Often the vocals or production changed, and so we'd have to adjust accordingly. There was a lot of files going back and forth between us, and so for security we'd only use USB drives or AirDrop.

No Deal

Tom Norris grew up in San Diego and learned to play piano and guitar. When in high school he joined a band called Allstar Weekend, which became quite successful. Just as the band was offered a big record deal, Norris quit, as one does.

"I felt a lot of pressure and stress," laughs Norris, "and I had gotten tired of touring. My parents are academics, and so I decided to go to university, UC Berkeley, where I studied linguistics. At the same time I got into making music on computers, and learning sound design in FL Studio on a PC. I used FL forever, and was using it for mixing up until the last three years when I switched to Mac and Ableton Live. I still remember people giving me weird looks as I set up my gaming PC at Westlake Studios, to mix a Hailee Steinfeld and Grey song!"

Norris started working with electronic duo Grey in 2015, with Zedd in 2016, and in 2017 began a collaboration with Skrillex that lasts to this day. "I've used a lot of DAWs over the years, and as a result I am pretty comfortable on any of them. If I get a session as a Pro Tools session, I will do the mix in Pro Tools. I have also done mixes in Logic and PreSonus Studio One. The reason why I ended up in Ableton was primarily because Sonny [Skrillex] is using it. We've been working together on almost every project for the last several years, so it made our workflow a lot easier. Also, Grey are roommates, and they use Ableton. Grimes also uses Ableton. For whatever reason, most people that I work with use it."

Today, Norris has two studios in LA, one in Skrillex's studio, and one at his home. "Both my studios have a Mac Pro with Ableton, a UA Apollo X8, and Focal SM9s in the studio at Sonny's place, and PMC Result6s at home. The PMCs are amazing. I also have a few pieces of outboard, like a Tube-Tech CL1B, a Neve 1073 to occasionally track things, and a Buzz Audio SOC 2.0 opto compressor, which I use a lot on my stereo bus. I also have an SSL G bus compressor, and I ended up using that on every song on the Gaga album, mainly because I was forbidden from using side-chain. I had to rely on hardware VCA compression to get the sound of pumping artifacts. I would push the kick or whatever pretty hard into the SSL to get that sound."

Look Mum, No Side-chain

Hang on a second... an album influenced by EDM, without side-chaining? Norris laughs again: "I know. Side-chain is the definitive sound of EDM, especially invisible side-chain, with super-clean and tight lows. But the thing is, Chromatica is a throwback dance album. We wanted to capture the '90s, and part of the challenge for me was: how do I do that authentically? To my knowledge, side-chain was possible in that era, but difficult to do. Burns and I discussed this early on, and we were like, 'Let's just not use side-chain, and see what happens.'

"Of course, I grew up with side-chain. The moment I learned about it, I started abusing it! It's kind of essential in modern records, because to maintain density, you need to side-chain to get rid of certain things in certain moments, so you don't lose headroom. Obviously, the same can be achieved with manual editing. But the goal wasn't to make an EDM album. We wanted it to sound more era-specific. The aim with Chromatica was to make it feel like '90s dance, but sound like it fits in 2020. I felt like our goal might have been similar to the spirit of '24K Magic' by Bruno Mars."

The way Norris talks about his work on Chromatica suggests that he was more involved than 'just' being a mixer. "Yes," agrees Norris, "I guess the best way to describe what I do is that I'm a finisher. I'm not involved in many purely hip-hop or trap things, and certainly my experience in dance music is a little different, in that people are more open to me trying stuff. You're often fighting 'demo-itis', but with the Gaga album there were several songs that I contributed additional production to. Recently, I've seen myself in this 'finisher' role more and more."

'Rain On Me'

'Rain On Me'. Written by Ariana Grande, Lady Gaga, Bloodpop, Boys Noize, Martin Bresso, Matthew Burns, Nija Charles & Rami Yacoub. Produced by Burns & Benjamin RiceNorris described his process of finishing 'Rain On Me'. "That song was the most involved on the album. By contrast, the deluxe version of the album has a song called 'Love Me Right' that took me roughly 40 minutes to mix. Burns' production was really great, so I didn't feel I needed to do much. But 'Rain On Me' went through at least two different productions. Tchami produced one, and then Burns, and the direction was chosen, which was influenced by French acts like Cassius and Stardust. But that was also then changed a lot. Burns and I ended up tweaking many of the sonics and I added a few sounds with his approval.

"I added an Omnisphere piano to emphasise the drop chords, and we experimented with the sub for a while. One of the main issues that kept arising was that it felt like there was not enough bass, which is a common problem in dance music, if the song is in a certain key. Some keys just don't have the greatest low-end response. So the question was: 'How can we still make this feel right, given its French touch, and still make sure it has weight?' The song is in C#m, which is usually too low to work in a house context.

Some of the compressors used on 'Rain On Me'. Tom Norris originally experimented with using the Alesis 3630 (top) to achieve a pumping sound without sidechaining, but eventually settled on the SSL G-series bus compressor (second from bottom)."I brought in my Alesis 3630, which I bought in the mid-'00s when I heard Daft Punk and Ed Banger were using it to get that pumping French sound. I sort of reversed-engineered 'One More Time' to see if that processing would work in the main riff of 'Rain On Me'. It didn't make it, but it helped me understand the general approach they might have had.

"But 'Rain On Me' was done over several months. Everyone was giving notes on it. Sometimes Gaga would come in and redo a line or a section of the song, so Ben had to re-edit and re-produce parts of his vocal mix, and I then had to work these in. Max Martin was giving us some suggestions a week before we had to submit the final version. Ultimately, work on the song started in November, and the final version was completed in March."

The Mix

Norris highlights some of the most interesting aspects, first elaborating on his mix process: "Some people start with all tracks at [minus infinity], and then they push up the kick drum, and then the snare, and the vocals, or whatever. But I typically like to listen with everything in together, and I try to get a balance as fast as I can with everything playing at once. I feel that the first 45 minutes I work on a song are the most productive, and beyond that I start losing objectivity and perspective. So I will typically first do volume adjustments and add whatever effects I want to hear, and then I will go through individual things and do spot adjustments.

"But I don't like to mix with tracks in isolation. I feel like I get the best results when I am hearing the gestalt of everything playing together. With this song I had everything playing at once, but I remember spending a lot of time adjusting the drums. The feeling of the off-beat hi-hat was especially important, because that is such a thing in house music. I definitely wanted to make sure I got that right.

"A lot of mixing the drums was just about level adjustment. On the drum bus I had the Plugin Alliance Black Box, the Slate Digital FG-X, and Oeksound's Spiff transient shaper. I also used the Ableton EQ8. It's really well designed, uses very little CPU, and I always switch on oversampling to reduce aliasing. I have a second drum bus, which has an ADA flanger and some filters which Burns and I added later on to add an effect in the pre-chorus. I really love Ableton's Audio Effect Racks, which allow me to create and save self-contained processing chains. Effect Racks are really the lifeblood of mixing in Ableton for me.

"For the drop bass, we tried several parts layered to get it feeling right. I added a Serum sub, there's an 808, and Burns also recorded a bass guitar. These ultimately were consolidated into one track for the most control. On the bass group, I have FabFilter Pro‑Q 3 on dynamic mode to reduce 70Hz a bit, and I used the Kush Audio AR-1 compressor to even out the signal. For the most part, the song has live instruments, including guitar, that are augmented by synths. I really love using the Ableton factory delay, on which I have a sample delay/Haas preset that I used on the guitars. There's also some low-end filtering in M-S in the drop section.

"At some point it was decided that the production was becoming too dense and turning into soup at points, so Burns and I spent a day going through to see what bits of the arrangement we could get rid of. For the synths, I didn't need to do much. I used the DJM filter during the pre-chorus to filter everything into the chorus. I also used a plug-in that I've been working on called Fresh AIR. I started developing this and another plug-in while at Henson just to get back into C++ programming. Fresh AIR is a brightening EQ, and it was really handy in adding smooth top end to parts that felt too dark. The other plug-in I made is a de-esser that works a bit differently from how traditional de-essers work. I had some ideas for what I wanted them to do, and thought it'd be fun to develop them myself!

"The keyboards also have Fresh AIR, FabFilter Pro‑MB, and Waves RVox on the Omnisphere piano that I added in the drop. I used Waves CLA Vocals on the high strings, with everything off besides stereo widening. I didn't do too much on the vocal chops, other than do some broad-spectrum resonance clean-up. I have the Oeksound Soothe and SoundToys MicroShift on those to widen things a bit. There were some volume automation moves on the vocal stems, but these were fundamentally minor adjustments because Ben had already done such great work on his end. I ended up using my de-esser on LG and Ariana's lead and background vocals just to dim the esses slightly, in response to all of the top end that was added throughout the mix.

"On the master bus I have an Effects Rack that I called Master Volume Ride. Ableton's macros are another great thing about their Audio Effects Racks, because you can create parameter thresholds for whatever knob you are adjusting. So the max values here are -1dB to +1dB, which means that I can do these huge moves with automation, but it is only changing things by a fraction of a decibel. Besides that, everything is going into a hardware SSL G bus compressor, using Ableton's External Audio Effect module."