Mix Masters: Mick Guzauski, March 1998MIXING SECRETS FROM THE TOP OF THE CHARTS 4/01/2001 1:42 PM Eastern
He’s the master mixer of smooth and soulful pop, with a list of Number One singles that’s longer than many engineers’ entire discography. Hit artists from Mariah Carey, Toni Braxton and Michael Bolton to Barbra Streisand, Boyz II Men, All 4 One and LeAnn Rimes seek him out to help craft the sound that keeps them out front and on top, and he always delivers. Success hasn’t gone to his head, though; Mick Guzauski is a genial, unassuming fellow. The polar opposite of jaded, with almost 30 years spent in the music recording business, he still evinces a sincere and almost childlike fascination with the various technical and artistic elements that comprise his chosen field. These days, Guzauski mixes primarily at his own Barking Doctor Studios about an hour out of Manhattan, where recently he’s been locked in with new projects for Eric Clapton, Lionel Richie and Monica.
Let’s take it from the top—how’d you get started in the business?
I was always interested in music and electronics, and although as a kid I had music lessons, I never believed that I’d be a very good player. But I was pretty good with electronics, so…
Did you build from electronics kits?
Kits, and I built a lot of my own stuff. I started recording as a hobby in the late ’60s, so of course there was no semipro gear—nothing affordable for a kid. Then, when I was 15, I got my first job repairing turntables and tape recorders for a local hi-fi store, and I started buying used and broken equipment and fixing it up. Then I built a mixer and started a studio in my parents’ basement; I was recording demos for local bands by the time I was a senior in high school.
You had a studio in your house even back then?
It was tiny, but in Rochester, New York, there were very few studios, and they were mostly in people’s basements. I did that for a couple of years after high school, then I went in with some other guys who had a commercial studio because my parents didn’t want bands in the house all the time. I was doing jingles, as well as those band demos, and I’d been doing it on weekends. But it got to the point, when I had a Scully 4-track machine, that we’d do tracking on weekends when my parents weren’t home, then do overdubbing and mixing on weeknights.
So you went into business with those other guys…
They did mostly industrial work, and I kept up the music end—more bands and jingles. There really wasn’t that much work in Rochester, so I did P.A. work, as well. I did P.A. for Chuck Mangione and ended up recording for him in the studio and also live in the early ’70s. Then, in ’75, he signed with A&M and was getting ready to do the album that became Chase the Clouds Away, with a live rhythm section and a 44-piece orchestra. He asked me, “Do you think you can handle it?” Of course, I’d never been in a real professional studio, and of course I didn’t think I could do it, but I said, “Yeah, sure.”
So I went out to L.A. with him and did the album; it came out good, and everybody was happy. After that, I went back and forth to L.A. for three years, in the studio with Chuck and also doing road work. I finally moved to L.A. in 1978. Now, 20 years later, I’m back in New York!
Your client base grew from Chuck’s albums—people heard your work and liked it.
One of the last records I did for Chuck was a pretty big hit called Feels So Good, so people had heard of me as an engineer, at least in instrumental jazz. I started getting a lot of calls to do that kind of album—you know how you get sort of pigeonholed. Meanwhile, I’d also gotten my first real job in L.A. as a tech at Larrabee. A friend who worked at Westlake called me up one day, and I went over to see the place—I ended up doing a Chuck Mangione project there, and I became close to the staff. One night, a band was in with James Newton Howard producing and Andy Johns engineering. Andy had double-booked himself, so Jim Fitzpatrick, Westlake’s head tech, needed an engineer and called me. The label’s chief engineer happened to be George Massenburg. I did a rough mix that he liked, and that’s how I got into working a lot in L.A. and went on to work for Earth Wind & Fire. Things happened faster after that; I was working at Conway Studios with Average White Band, and I met Peter Bunetta and Rick Chudakoff. Through them I started working with Smokey Robinson, Michael Bolton, Kenny G and Walter Afanasieff—that’s really the chain of events. David Foster, who I do a lot of work for now, I actually met back in Rochester where I did a demo for him in ’73 or ’74 when he was in a band called Skylark, who were managed by someone who lived in Rochester.
You never went the runner/assistant route?
I never assisted; I came more from a technical aspect. I’m not a designer like Massenburg, but I built stuff for my home studio because I wanted to understand how it worked. I had time on my hands in the early years, and at one time or another, I designed and built every part of the analog chain. Not a whole console, but one channel each of record, tape and play, so I understood how everything worked from the mic pre to the power amp. Analog, of course. I started getting really busy before digital technology became prevalent, so I haven’t had time to learn as much as I should about it. I want to, though; I want to get a DSP development kit and see how it all works.
Your records always sound so warm that I was surprised to learn you work mostly with digital.
I do work pretty much all digitally—it just makes sense because we have the digital final product of a CD. While it’s true that analog does tend to warm things up somewhat with tube gear and such, I believe there’s no reason digital can’t be warm. I never really have much problem with that. I mix to a DA-88 through a Rane Paqrat. The Paqrat takes a 24-bit digital word and splits it onto two 16-bit tracks of an 8-track machine; you can record 24-bit stereo audio using four tracks of a DA-88, and it will reassemble the 24-bit word on playback.
The system I use now is the AT&T digital mixer core for my SSL console, with a Sony 3348 multitrack. I have an analog multitrack for when people send me analog tapes, but I usually transfer them to digital, then we stay in the digital domain all the way through. I also have a lot of digital outboard gear—TC5000, Valley Compressor, Weiss Equalizer, Eventide DSP4000 and H3000, Lexicon 300s and a Sony V77—a lot of gear that I can go in and out of and keep everything in the digital domain. I do also use analog—a GML EQ and compressor, and I have a couple of old Eventide SP2016s and an EMT140 plate that I use in combination a lot on vocals; I think that’s where some of the warmth in my sound comes from. The 2016 has a nice high-frequency sizzle and definition in the reverb, and the plate has that big, warm sound behind it, because, although the EMT is transistor, not tube, it’s still a mechanical device.
The vocal ambiences on your records are always very clear; you may have a lot of reverb going on in a ballad but it doesn’t get in the way.
I always try to make sure that the low end of the echo isn’t in the way of anything. Often, I’ll EQ the reverb to attenuate lows, so that when it rings in the lower register of the voice, it doesn’t cloud any instruments. Usually, I’ll roll off some low-mids, around 200 or 300 Hz. I’m very careful that I don’t lose richness, but most of the main keyboards and pads will be right in that same area, and it can be a balancing act to keep the vocal clear. I also work very hard on the high-end detail in the vocal to make sure that it’s very present—not sibilant, but with enough diction so that you can hear the vocal in the track without putting it incredibly up front.
Are there particular compressors you like to use on vocals?
I’m using UREI LA22s a lot; they were only made for a couple of years. I like their flexibility. In addition to variable attack and release time, the detector is variable from peak to average. Also, you can use it as a frequency-selective compressor—not in the way that a sidechain is frequency-selective, but like a dynamic EQ that will dip a certain frequency. If you have a vocal that at certain levels and frequencies jumps out and gets harsh, you can use a channel of it and actually detect the threshold of that jump, then dip that frequency.
In general, do you vary the attack and release times a lot on compressors?
Actually, for vocals I’ll usually leave them set pretty much the same. What seems to work is about a 3-to-1 ratio, with a 20 to 40-millisecond attack time and a 100 to 300-millisecond release. I don’t just leave it set that way for every voice, of course, but I find for most applications, that range works. I usually adjust attack and release a lot more for instruments than on vocals—the envelope of the voice in general is usually close enough so that when a compressor is set this way, it will attack pretty well without pumping. With a slower attack time, you’ll hear the vocal attack and then you’ll hear the compressor grab—it will be audible. If the attack is faster, you sometimes don’t get the transient of a word that’s being emphasized. So in general these settings work. Once in a while, if something is sung very staccato, I’ll have to speed up the release.
It sounds like you ride reverbs and delays a lot during the mix.
Yes, usually I have a couple of automated channels on the console that are sends to different delays and reverbs, mainly for lead vocals and lead instruments. Sometimes it’s an effect to really make a word swim, but more often, the reverb that would sound plenty wet in a sparse part of the track will be too dry as the track builds dynamically. So, often I’ll ride the reverb and delay as the track builds.
You’ve done a lot of group lead vocals like Boyz II Men and All 4 One. How do you deal with so many vocals?
You’ll usually find when you place each individual vocal in the track that certain frequencies will pop out in certain areas, while in other areas they won’t be loud enough. When they are soloed and un-EQ’d, they sound good, but there may be little spikes that stick out in relationship to the track. So I’ll EQ each vocal individually to be smooth, generally with parametric equalizers with the bandwidth set pretty narrow. I’ll boost a frequency a lot to tune right to it, then cut it back to be smooth and in context with the rest of the spectrum. Sometimes I’ll split a vocal on two channels and have a different EQ for verse and chorus. And sometimes frequency-selective compression will help. If you have a vocal EQ’d pretty bright that sounds nice when the vocalist is singing soft but which gets a little nasty when they sing harder, that kind of compressor can automatically ride those changes.
If I’m going to cut a frequency back that’s sticking out in the mix, I’ll tune it narrow and sharp to find the exact center frequency and boost it to find it, then broaden it up and cut it back as much as I need to. Most of the boosting I’d do is very high-shelf, broadband, so you don’t hear one note jump out, unless, of course, something is deficient in one small area.
And with backgrounds?
With BGs, I’ll EQ for smooth response; I don’t compress them that much. Usually, I get them in stereo pairs per part—one part that’s a chord, and a counterpart to that. Sometimes the parts are individual. Then I’ll bus them to a pair of channels so that I can work on them as a whole. I’m always listening for the chords. They may sound perfectly balanced soloed, but something may get lost when it’s in the track. I have to say, though, most of the vocals I get are very high-quality, very well-recorded. Well-balanced within themselves and also with the track.
Do you use your console EQ primarily?
Yes, mostly the console EQ, except usually on lead vocals I use the GML parametric. My chain for that tends to be out of the converter to the GML EQ, then the compressor, the de-esser and the console EQ. That way, if I need to cut back on something, I use the GML, and the compressor is reacting to a fairly smooth signal. Boosting I’ll do on the console, post compression. So usually I have EQ both pre- and post-compression—cut before compression, boost after.
With the AT&T system, the console is switchable per channel between EQ that emulates both SSL E and G Series and Neve VR curves. Usually, I keep it on SSL E, but on acoustic guitars, strings and some high percussion, I’ll go to the Neve VR curve because the characteristic of the high shelf is very airy—it has a little dip before the shelf starts and then sort of comes up a little steeper and gives you air without getting harsh.
I think of you as a bass expert. Your bottom end is present and big but never loose and boomy.
Well, with some projects I work on, the bass is in such a low register that there’s no attack with it, but it works musically. That’s where I’m very careful setting the attack time on the compressor. I’ll listen to the bass in the track and check the attack of each note. Then, if you slow down the attack of the compressor and compress the track a little, the compressor will actually add some attack because the attack transient of the bass is getting through before the compressor reduces the gain on it. That’s a trick I use a lot—to use a compressor with a slow attack and then to adjust the amount of compression so that there’s enough of the attack of the note that gets through to define where it is in the track. I also EQ it to make the whole register that the bass is playing in, the whole spectrum, sound smooth with the track. That doesn’t always mean that every note is perfectly even, because something else might be covering the bass notes in some areas. You want to tailor it so it sounds even in the track.
The important thing is, I don’t listen to individual tracks alone too much. I’ll usually listen to a rough mix first if the client sends it, or else I’ll put up all the tracks to hear what the song is about. Then I’ll go through the tracks and EQ and set them pretty much where I think they’re going to go, and then I’ll put everything in and fine-tune.
Like mixing a live band.
It is probably a carryover from when I used to do live sound work.
What compressors do you like on bass?
Most of the time on bass I seem to use the Empirical Systems Distressor because it is so adjustable. I’ll also sometimes use an older limiter like the LA-2A or the clone made by DeMaria—it’s not very variable, but it has a very complementary attack and release time by itself. That’s another optical; I use a lot of the optical compressors. Even though they don’t have variable attack and release, their normal characteristics can work very well.
Do you have to do a lot of mixes for each song? Vocal up, down, etc.
Not so much; a mix, then a vocal up, a vocal down, so if later a word or a phrase isn’t quite the right level, you can cut them in. I’m also mixing in stems a lot now, where I’ll turn the bus compressor off, because of course that compressor in the final bus reacts to everything, and I’ll do on DA-88 a pair for the track, a pair for BGs, a pair for leads, a pair for orchestration—four different stereo elements of the mix with their effects. That can be combined and sent to a bus compressor if we need to change balances without redoing a whole mix.
I’m also using Pro Tools a lot for that. I should mention that’s now one of the most important things in my studio. I have Pro Tools 24 with 24-bit resolution, so I can put a whole mix in it without really losing anything. I use it for editing mixes, for flying and moving stuff around. I also use the plug-ins with automated EQ. If a vocal comp needs different EQ dynamically as the song goes on, I’ll put the vocal in Pro Tools and let it run in sync.
There’s also an auto-tune, another plug-in which is the greatest thing. You tell it what key the song is in and how fast and critical you want it to be about pitch. I’ll just copy the vocal track into Pro Tools. That’s another thing I really like about working digitally: You can copy back and forth, and as long as the clocking of the system is stable, you don’t really have any loss. That’s a point that should be made about digital: Generation loss in digital is nonexistent only as long as the clocking of the system is stable, all the connections and the wiring are good, and you are not getting noise errors into the chain.
Your console is a 56-input SSL G Series.
Yes. The AT&T digital mixer core has a little card that goes on each channel of the SSL, each multiplexer board, and it picks up the recall information there and sends it to a computer and the AT&T DSP, which is actually just modeling what the whole console does. So the SSL isn’t passing any audio at all; the recall section is all that’s running. When the AT&T is operating, it forces the recall circuitry on all the time, so that whenever you are turning a knob or pushing a button, it knows what you’ve done. And then, of course, it reads the automation for the faders and mutes.
You must use your ISDN lines a lot.
Our system is a Dolby Fax through EDNet, and definitely ISDN lines are one of the reasons that I’m able to have a home studio and to do most of my work here. Being this far from the city, producers don’t always want to come, and also I still have a lot of clients in L.A. For the most part, people don’t have the time to get on a plane and come here. This way, I can do the mix and then the clients can either listen at home or at a studio and we just tweak it together for the last couple of hours. Then we’re done.
Which other engineers’ work do you like?
Off the top of my head? I love a lot of the stuff Bob Clearmountain does; I think he’s a master at getting really interesting spatial things happening. I love Elliott Scheiner, the precision and fidelity of his work. There’s a naturalness to it, a natural ambience like everybody’s right there in the room that’s very good. I’ve always loved his Steely Dan records, and more recently, Fleetwood Mac’s The Dance is just great. I love Swedien, Massenburg, Al Schmitt—there’s too many excellent engineers to mention them all. There are a lot of great people in this business.
How about some thoughts on the philosophy of mixing?
Well, I think it really is a new profession now, because it’s only in the last few years that records have been made that are almost totally synthesized. So being a recording and mixing engineer are almost two different things now. There are still engineers who do both, guys like Al Schmitt and Elliott Scheiner, who are great engineers and great mixers, and there is the type of music that really needs somebody to do both and to follow a project through—those projects with mostly live instruments, where so much of the mix is created in the recording.
But being a mixer, like me or Bob or Jon Gass, and all the guys that people send their tapes to, really is a pretty new profession. We’re creating an ambience for a song, sort of like a photographer does for a picture, because a lot of pop music is mostly synthesized and not recorded with its own particular ambience. So, different mixers interpret the music differently. It really is a different set of responsibilities.
One last thing. Tell us your secret for getting all those Number One singles.
Honestly, I have no idea how it happened, but I’m sure glad that it did! Seriously, though, as an engineer or mixer, there would be nothing at all to do without great artists and producers. So that’s how it happened. I was lucky enough to hook up with great people, and I’m really thankful to all of them.