Inside Track : Recording Evanescence's 'What You Want'

The multitracks for Evanescence's third album were so big that they required two maxed-out Pro Tools rigs to play back!

The way the music business works,” says Randy Staub, "is that you get pigeonholed as an artist, a writer, a producer, an engineer and as a mixer. If you've had some success with heavy rock, like I had, people will naturally think that that's all you do. I do like rock music, but I don't like music because of its genre. I like it because it's good. It can be extremely heavy, or it can be Hank Williams, or 50 Cent. I love all kinds of music, but sometimes, when I'm 13 tracks into mixing a heavy rock album, I find myself wishing that I was mixing a girl singer with an acoustic guitar!”

While Randy Staub's sentiment is understandable, it's also understandable that he's regarded as a living legend in the world of rock, having worked with Mötley Crüe, Nickelback, Metallica, Bon Jovi, Iggy Pop, Alice In Chains, Bryan Adams, Hinder, Lostprophets, Evanescence and many more.

Hailing from the town of Prince George in British Columbia, Staub always wanted to be an engineer. He recalls, "I always liked the sound of records and the technical aspect appealed to me, so to become an engineer was the obvious thing.” After leaving high school in the late '70s, the Canadian attended a summer recording course in Rochester, New York, and spent a while doing live sound before being employed as an engineer at Phase One Studios in Toronto. It was there that he met fellow Canadian Bob Ezrin, who recommended him for a job at A&M Studios in Los Angeles. Staub spent three years at A&M, but was eventually persuaded by another top Canadian producer, Bob Rock, to return to Canada, to work in the latter's Vancouver studio.

The first record Rock and Staub did together was Mötley Crüe's Dr Feelgood (1989). Two years later they worked on Metallica's eponymously titled album, also known as the Black Album, which went 15 times platinum in the US. Unsurprisingly, Staub has, as he says, "worked non-stop since”. Since Rock moved to Hawaii in the mid-'90s, Staub has mostly worked out of Bryan Adams' The Warehouse studio in Vancouver. Until 2001, his credits were fairly evenly divided between engineering and mixing, but for the last decade, 95 percent of Staub's work has been mixing. "It was a conscious decision to focus on mixing,” Staub explains. "It has always been the thing that I enjoy the most.”

Evans Above

Randy Staub mixed the Evanescence album on the SSL 4072 in Warehouse Studio 1.

One of Staub's recent high-profile projects was his mix of the entire third album by the American goth metal group Evanescence, including lead single 'What You Want'. The disc, simply titled Evanescence, ticks all the boxes that characterise much of Staub's work, with heavily distorted rhythm guitars, monolithic bass and drums, and an in-your-face sound image that is very, very loud. Singer Amy Lee's dramatic voice and crystal-clear acoustic piano (she's classically trained) are an essential part of the band's identity, and extensive string arrangements form the icing on the cake. Though it has yet to match sales of the band's debut album, Fallen (2003, 17 million worldwide sales) or its follow-up The Open Door (2006, five million), it reached the top spot in the US, and has so far peaked at number four in the UK.

The band began work on the album early in 2009 with producer Steve Lillywhite (U2, Peter Gabriel, Talking Heads), but eventually shelved these recordings, and a year later switched to producer Nick Raskulinecz (Foo Fighters, Alice In Chains, Rush). Raskulinecz and Staub had earlier both worked on Stone Sour's Audio Secrecy (2010) and Alice In Chains' Black Gives Way To Blue (2009), amongst other things, experiences that set the tone for a fruitful working relationship on Evanescence.

"Nick and I have done a few records together, which have generally been big rock records, and that was definitely the approach here,” says Staub. "I always tell people that you have to create the sound of the record that you want right from the beginning, so that when the mixer first pushes up the fader, the record is there. Everything sounds the way it should, the arrangements are good, the emotion is there, the balance is there, and as a mixer your job is to take it to another level. You pump it up and make it sound larger than life. This rather than trying to create something that wasn't there to begin with. Luckily Nick works like that. His stuff has its character recorded into it, so I don't need to mix it into it. The arrangements are great, and Amy's vocals are powerful and recognisable and tell the whole story of each song right from the start. So it was a matter of making her vocals sound even more powerful and making the drums sound bigger and making sure that the strings could be heard. There's lots of everything, and it all had to sound big!

"I mixed 16 songs over five weeks during the Summer of 2011. So on average we mixed a song every two days. Generally, my process is to work on a song and get it to sound great during the first day, leave it overnight, listen with fresh ears the next morning, make some changes, and then spend the afternoon printing all the different versions. One thing that can take time is waiting for feedback from the artist and producer, if they aren't there. I'll post mixes for them online, and hopefully they'll get back to me right away. But it can take an hour or a day before people get back to me with comments. I then make some changes, and post the updated mixes, and then wait again for comments.

"I like to talk with the producer and artist before I begin mixing, to get as clear a picture as possible of what they want to achieve, and for the same reason I like to hear the rough mixes, because people will have spent a lot of time on them, and they give me a basic idea of the kind of balance and perspective that the artist and producer have in mind. Zach Blackstone, my assistant, and I generally work by ourselves here in Vancouver, but for these sessions Nick and Amy came over for a couple of weeks, and I really enjoy that, especially when it's at the beginning of an album mix. When you're doing an album it can take a few songs to find where the album is going to sit, sonically, thematically, musically. Once the first two or three mixes are done, you know what ballpark you're in. You hopefully have a roadmap and you know where you're heading in terms of a vision for the album.”

'What You Want'

• Written by Amy Lee, Terry Balsamo, Troy McLawhorn, Tim McCord, Will Hunt
• Produced by Nick Raskulinecz

Randy Staub: "We mixed this song first — the record companies always want to begin with the first single, even though, as I mentioned before, it can take one or two songs to know where you're headed. This [Pro Tools] session was very well-organised, Nick is really good at that. In fact, we've recently been getting more and more sessions that are well-organised and clearly labelled. Enough engineers and mixers appear to have been impressing on people that it's important to send clear, clean sessions. We don't need Playlists, we don't need stuff hidden. When I open a session, everything that the artist and producer want in there should come up immediately, so I can see it straight away. One issue with this session was that we needed two Pro Tools rigs to run it! Apparently, unlimited tracks is not enough any more. The session was at 96/24, and at 96k you can max out a Pro Tools system.This composite Pro Tools screenshot shows the session on the 'A' rig. From top, the tracks are colour-coded purple (drums), green (guitars), grey (bass), and red (vocals). The grey and blue tracks at the bottom are auxiliary tracks containing send effect plug-ins.

"Zach gets in the studio before I do, and he will organise a session in the way that I like — I've been doing sessions the same way since the beginning of time. He'll colour-code everything, clean up tracks and adjust levels, making sure everything is within reason and nothing will be distorting the console when it comes up. Because of the amount of tracks on 'What You Want' [around 150] he combined some of the tracks, and I also had to premix things before they go into the desk. I have a 72-input console, but if there are nine guitars all playing the same part, they will usually come up on the same channel on the desk, and the strings will tend to come up on two faders, although if there's a string part that needs its own treatment, it'll come up on a separate set of faders.

"It is a big session, with about 23 tracks for the drums alone, 20 vocal tracks, and 72 string tracks, but the arrangement is the key to it working. That's often forgotten these days. Like if the bass is very low, as in this track, you can change the voicings of the rhythm guitars to make them sound lower as well. With strings, particularly in rock music, it is all about the parts that they are playing. It's so easy to have strings just playing chord changes, kind of like a pad, and if it's a pretty aggressive rock track, the strings will just blend in and you won't hear them. The fact that you can hear the strings in this song is mostly down to the arrangement. They are well arranged and all the individual strings play good parts.

"When I begin a mix, I will normally push all the faders up right at the start, and get a rough balance of everything. I'll then concentrate on the drums. I won't switch the instruments and the vocals out, I simply turn them down, so I can still hear them in the background and hear the drums in context. I normally begin with the top drum mics, like overheads, hi-hats, toms, rooms, and then I'll blend in the kick and the snare. The reason is that a lot of the sound of the drums is in the top mics. It's fairly easy to get a good sound from the close mics on the kick and snare; it really is the sound of the whole drums that's trickier. Is the hi-hat in phase or out of phase, is the second floor-tom out of phase, and if so, why? Are the room mics in phase with the overheads? On occasion I might nudge stuff in time. Once I'm happy with the drums, I'll put in the bass, make them fit, make sure all bass tracks are in phase, maybe go back to the drums, and so on. Once I have the drums and bass the way I like, I go to the vocals, and after that I'll blend in the other things. But the bass, drums and vocals are the foundation.”

Drums: Waves SSL EQ, desk EQ and compression, AMS reverb, Pultec EQP1A, GML 8200, SSL G384.

"Everything you see in the session came from them [ie. Raskulinecz and the band], apart from a couple of kick samples that I added, which I like to do, because you can EQ them very radically. They had an NS10 [a Yamaha driver used as a microphone] going to record the bass drum, and also added some tom samples. 'Kicksnare Compex' at the bottom of the screenshot is something they did, probably a mic between the kick and snare, on which they added a Compex limiter. It sounds very compressed, very trashy. The volume rides on the toms were also done by them — I would have done them if they hadn't. I put the SSL channel plug-in EQ on the tom tracks, and one of the snare mics, I think because I wanted these tracks to sound more similar. The five room tracks gave most of the space to the drums, and I had them coming up on a stereo pair on the desk — you can see the desk channel routing in the I/O column. I also had a little bit of AMS reverb on the drums, but not much. I used desk compression and EQ on the drum channels, and also some outboard EQ like the Pultec EQP1A and GML 8200 on the kick and the snare channels.

"On the console, I sent each small drum fader to what I call a drum compression group, on which I had an SSL G384 compressor and the GML 8200 EQ. Most of all, I'd send the kick and snare there, but also, in lesser amounts, the overhead and room mics. The amount of compression really varies per track — sometimes it's quite extreme, sometimes it's very little. On this song, I had a pretty solid amount of compression and EQ, and I'd blend the compression group in with the original channels. It's something that Bob [Rock] and I started doing years ago, and that's now pretty standard. I also often send some of the bass to the drum compression group. It is intended to make the drums and bass sound punchy, and larger than they really are. Part of mixing rock music is to get more excitement in a track than really is there, and compression seems to do that. Again, it makes it sound larger than life.”

Guitars: Neve 1073, desk EQ and compression, SRS Wow Thing.

"The guitars were underneath the drums in the session. There are two times four tracks of rhythm guitars, each representing one part that was played via two amps and each amp was recorded with two mics. One set of four, played by Troy [McLawhorn], came up on channels 9 and 10 and was panned left. The other, played by Terry [Balsamo] came up on channels 11 and 12 and was panned right. You'll notice the volume automation on Terry's guitars, which was to take out some of the noise in the quiet bits between the chords. There were no plug-ins on the guitars, just outboard. I had the Neve 1073 EQ and console EQ and compression. I don't generally use a lot of compression on heavy guitars, because the big amplifiers generally have compression and distortion built in anyway. If the sound is a bit muddy and it's a very choppy part, I may use compression to get some control over the lower notes. There's also a single track of chorus guitar that came up on channel 13 and that I simply EQ'ed to make it fit.

"One effect that I often use on guitars is a small effect box called the Wow Thing that I bought for $20 many years ago.Secret weapon: the SRS Wow Thing. I'm letting my secrets out here! At the time, I was listening to music on a computer at home, this was maybe in the mid-to-late '90s, and it sounded really good and super wide. I wondered how it was done, and when I dug into the computer I found this piece of software made by SRS Labs called the Wow Factor. The higher you pushed the slider, the wider the sound got. It really is just a phasing program that makes stuff sound wider, to the point that it may sound outside your speakers. So I went to their web site and bought this small hardware box that cost 20 bucks, and I found that it sounds great on guitars. I will send them to the box using the small faders and spread them out left and right. SRS also make professional spatial effects devices, and I do have one of them, but I find that this small box still works the best for me.”

Bass: desk compression and EQ, Pultec EQ, Focusrite Red 3.

"'Taurus Moog' immediately below the guitars is a low pad-type sound, and 'Audio 1' and '2' are two brief vocal parts on which I put the [Waves] L1 and the [Sound Toys] Filter Freak. It's an effect vocal that appears in the middle of the third verse. Below that are five bass tracks, one from an Ampeg amp, one DI, one from a sub cabinet, and the JMP1 and Sans tracks have a distorted bass sound, using the Sansamp. I had the Ampeg track on channel 20, and grouped the DI and sub on channel 21 and the distorted tracks on channel 22. I used desk compression and EQ on these. If I did use outboard on the bass, it would have been fairly minimal, perhaps some Pultec EQ on the DI and sub, and if I did use outboard compression it would have been from the Focusrite Red. But these were pretty solid bass tracks and I didn't need to make them nice and solid.”

Vocals: Waves SSL Channel and Renaissance Compressor, Avid Lo-Fi and Revibe, Line 6 Echo Farm, Retro 176, Dbx 902, GML 8200, Eventide DSP4000 and H3000, Yamaha D5000, Lexicon PCM42, Urei LA2A.

"The first three vocal tracks — 'Verse Vocal', 'Chorus Vocal' and 'Bridge Vocal' — are the lead vocals. You'll notice that the chorus and bridge vocals both come up on channel 24, while the verse vocal comes up on channel 23. I wanted to give the verse vocal a different treatment, something that appears to be fairly common nowadays. The verse vocals are often sung with different volume and energy and require different treatment. The plug-ins on all the vocals were already there, and I left them. When a session has plug-ins, I tend to use them, because I assume that the sound they give is what's intended. The verse vocal has the SSL Channel plug-in on the insert, which I rarely use because I have an SSL console. 'R' is the Renaissance compressor, and you'll notice that it's on all the vocals. 'L' was the Lo-Fi plug-in, to give the verse vocals a little bit of edge.

"The sends go to reverb and delay effect tracks that also were already on the session. I generally use my own reverbs and delays, but they were happy with the sound that they had going, so I used that as well. They'd spent a lot of time and effort to get something they liked, so I used it and hopefully improved on it. [Send] 2 has the Revibe reverb and 3 and 4 have Echo Farm delays. You can see these effect tracks at the bottom of this Pro Tools session. They came up on channels 39-40 on the desk. Below these three effect tracks, you can see my two mono Echo Farm and two stereo Revibe reverb tracks, which were coming up on channels 47 and 48 and 49-50 respectively. I recall that after I mixed this song, ie. when mixing the other songs, rather than send the effect from Pro Tools to the desk, I sent the signal from the desk to Pro Tools, the reason being that if I made a volume adjustment on the desk, it wouldn't change the level of the reverb.

"I really like the sound of Revibe, and I use it like a piece of outboard gear. Whether something is in an outboard rack or in Pro Tools, it's still a digital reverb. Other rack outboard gear that I used on vocals included the Retro 176 Limiting Amplifier as my main vocal compressor, the Dbx 902 de-esser and the GML 8200 EQ, and I may have used a GML compressor as well. As always, I would have used console EQ and maybe some console compression. The Retro is my own, and I really like using it on vocals. I also may have used the Eventide DSP4000 and H3000 for chorusing and delays and the Yamaha D5000 and PCM42 for delays on the vocals. I'll treat the background vocals differently than the lead vocals, to try and separate them a little bit, and may have had a 'Dual 910' program on the H3000, or different chorusing and delays. All EQ and compression on the BVs would have been done on the console, with a little bit of LA2A as well.”

Keyboards: desk EQ, Pye compressor.

"Underneath the vocals is a 'Piano Effects Print' track, which was the only piano track that I had. It had the effects that they wanted, so I did nothing else to it, other than desk EQ it to make it sit in the track. The other keyboards are in the other Pro Tools session, at the top and bottom of that session. At the top are some pretty weird-sounding synth overdubs, so they have their tone and sound built in, and I didn't do much to them, other than some desk EQ to make them fit. There's also a loop, on which I had a Pye compressor. At the bottom are some more synth sounds, arpeggiated ones, and bells, and stuff, and again that was purely a question of EQ'ing them to make them fit in the track.”

Strings: Avid Revibe, desk EQ.The 'B' rig Pro Tools session contained synth parts and loops (top and bottom) and four takes of a small string section.

"All the strings came up on 41-42. I would balance and pan the strings in Pro Tools, with the violins hard left, the violas soft left, the cellos on the right and the basses in the middle. There was a mono room sound on all string takes that I did not use. Instead I sent all strings to a Revibe effect track in Pro Tools, that also comes up on 41-42. The Revibe has a medium dense church setting. Takes one and two came up on subgroup 21-22 within Pro Tools, and take three is a different part that was routed to subgroup 25-26 while take four went to busses 29-30. You'll notice that there's a big volume ride in the middle of the bridge on Take 3. Take three is panned hard left and take four is panned hard right. Because take three and take four play different parts from takes one and two, I wanted them to stick out a bit more. In addition to this, I only used EQ from the console on the strings.”

Master bus: SSL Quad Compressor, Sontec MES 432.

"I mixed to another Pro Tools rig that was also at 96/24, via the Apogee PSX100 [A-D converter]. I put an SSL Quad Compressor over the stereo mix — it is built into the console — and after that the master fader went to the Sontec MES 432, which is one of the best EQs ever made. The amount of compression I use on the Quad can be from almost none to 8-10dB at 4:1. In this case it was 2:1, with a fairly slow attack and release. I then sent off the full-bandwith file to the mastering engineer. What happens in mastering is, generally speaking, out of my hands, it's between the mastering engineer and the producer and artist. Records have become too loud, and I don't like it too loud. You can only go so loud until you reach a point where records don't sound very good any more. They become too linear, are fatiguing to listen to, and will distort on consumer electronics, which can't take the level. I think Nick is on the same page here. Yeah, this album is loud, but I don't think it's the loudest thing out there. We definitely try to impress on the mastering engineer and artist that the track needs to have some dynamic range, so that people can turn it up and it will still sound good.”