You are here

Vocal Fixes

Modern Vocal Processing In Practice
Published October 2003
By Dan Daley

Vocal Fixes

A new generation of vocal processors, and a new generation of engineers, confront the possibilities and ethics of digital vocal manipulation.

Like many CD covers these days, Nashville recording artist Allison Moorer's latest has a sticker on the jewel box. However, Moorer's does not warn parents of explicit lyrics; rather, it advises that it was recorded and mixed in a manner that's becoming increasingly rare. The sticker on Moorer's Miss Fortune LP states boldly: 'Absolutely no vocal tuning or pitch-correction was used in the making of this record.' Moorer's producer, RS Field, observes, "Vocal tuning is contributing to the Milli Vanilli-fication of modern music. [Putting the sticker on the record] was sort of our little freak flag."

Has the pursuit of perfection gone overboard? Listening to the radio, it might seem as though the recorded vocal is experiencing something like what the guitar underwent in the 1970s, when a proliferation of stomp boxes spiced guitars with fuzz, flanging, wah-wah, phase-shifting, envelope filtering and other processing flavours. The result set on edge the teeth of many purists. On the other hand, effecting vocals has been an artistic pursuit throughout the history of recording, from running a speaker and microphone into a tiled hallway to create reverb, to using springs and plates and finally digital reverbs. The James Gang's 'Rocky Mountain Way' and Peter Frampton's 'Show Me The Way' famously used voice boxes that allowed the guitar to modulate the vocal, and there were plenty of other late-night processing experiments over the years, some of which survived morning-after scrutiny to make it onto records. And then there were The Chipmunks...

Surrounded By Plug-Ins

The vocal is the one element of a recording that can't yet be emulated by a computer, but it's increasingly encircled by a rapidly expanding pool of plug-in processors, the sheer availability of which seems to have precipitated a new era in vocal effectation. "[Plug-in processors] are like pretzels on a bar," says Nashville engineer/producer Bob Bullock, whose credits include the Tubes, Reba McEntire and Shania Twain. "They're within easy reach and you can't eat just one." The watershed of vocal pitch effectation may have been Cher's smash single 'Believe', which in 1998 put unbridled digital vocal processing on the front page. Since then, artists from Lenny Kravitz to Madonna have created purposely processed vocals that cross the border into the domain of distortion.

Nashville engineer/producer Bob Bullock: "[Plug-in processors] are like pretzels on a bar. They're within easy reach and you can't eat just one."Nashville engineer/producer Bob Bullock: "[Plug-in processors] are like pretzels on a bar. They're within easy reach and you can't eat just one."The focus of what has turned from a purely technical pursuit into an ethical debate — one that's played out on the pages of general-interest media as well as in the audio trades — is not so much the digital audio workstation as it is the third-party plug-in from Antares Audio Technologies, Auto-Tune. The device — which had a parallel hardware version used live by Cher — has been both praised and pilloried by audio professionals and music aficionados alike, with plenty of both on either side of the debate. Michael Logue, marketing director for Antares, in Scott's Valley, California, relishes the discourse. "Sure, we've been vilified," he says affably. "But I then ask people, when you use a word processor, do you use a spell-checker? It's really no different than that."

Logue says that Cher's 'Believe' was a watershed not only for the application of automatic pitch-correction, but also for its users coming out of the tech closet. "The song made it OK to admit that you used it," he says. "That has really opened up the artistic possibilities of Auto-Tune. We now see our hardware versions in live sound racks on tour with arena-sized acts. It's also used on the broadcast feed of the Grand Ole Opry. It's used far more than people ever realise."

(Just for the record, Antares chief scientist, Dr. Andy Hildebrand, is an award-winning exploratory geophysicist. Auto-Tune's algorithms were based on work he had done in developing systems that do real-time geophysical seismic data interpretations. He made his mark some years ago, as Exxon's youngest-ever vice president, developing a crawling computer, dubbed 'The Pig', which would travel the length of oil pipelines using sound to check for structural integrity. That product, ironically enough, was based on research he did for another product — this when he went back to university to learn music after already having achieved several PhDs — called Infinity, which was the looping processor used in many synthesizers since the early 1980s.)

But, in a larger sense, DAWs have become synonymous with a quest for perfection that can leave music soulless — 'Pro Tools' has become a generic, in much the same way that 'Xerox' has come to be a trope for any copying machine, and is becoming emblematic of the antithesis of reality. "This is definitely our anti-Pro Tools record," Pearl Jam drummer Matt Cameron recently told a reporter on the release of the band's last album Riot Act. "It's more interesting hearing musicians in a room playing hard, with the tempo fluctuating slightly as the band heats up. Perfection is boring."

Geoff Foster, a veteran engineer at Air Lyndhurst, was recently quoted in print as saying, "To some degree, Cher's got a lot to answer for... Sure, [Believe] was an extreme example and meant to be an effect, but the public bought it — literally. From a pop standpoint the public said, 'We don't care what our vocals sound like'. But in a way, that was a testament to the fact that the musicality of recording can easily become secondary to the technology. Cher can actually sing, but the current generation of pop stars have been given a mandate that they don't need to."

Foster goes on to say that it's not just the artists, but a new generation of audio professionals who have bought into vocal tuning as a de facto standard. "There's a whole generation of engineers who have grown up thinking that the first thing you do when you get into a recording session is fire up Pro Tools ready to do repair work," he said in the same published interview.

Slacker Sonics

Josh Binder, engineer/programmer for producers such as Dave Reitzas and Walter Afanasieff: "[Auto-Tune] is not there to fix the vocal; it's there to be part of the vocal sound."Josh Binder, engineer/programmer for producers such as Dave Reitzas and Walter Afanasieff: "[Auto-Tune] is not there to fix the vocal; it's there to be part of the vocal sound."Not all engineers see it quite that way, though. Josh Binder is a twenty-something engineer/programmer in Los Angeles. He works regularly with producers such as Dave Reitzas and Walter Afanasieff, and his work appears on recordings by Kenny G, Barbra Streisand and Charlotte Church. He speaks of the digital domain as though he grew up in it and he's an expert on Auto-Tune, but adds that he has broadened his appetite to programs such as Wave Mechanics' Pitch Doctor — his take there: "It doesn't change the sound of the vocal like Auto-Tune does, but it has a massive delay you need to offset for, like 80 milliseconds," he says — and TC Electronic's Intonator. In fact, Binder continues, automatic pitch-correction is moving beyond its original mandate of fixing vocals and is becoming part of the sound of vocals.

"When you're working with a great singer whose pitch is right on, you can still apply Auto-Tune," he says. "I'll throw a chromatic Auto-Tune [patch] onto the vocal with a kind of mellow responsiveness level, which gives it a nice chorus/flanging effect. I'll print the effect to a separate track and then paste it into the comped vocal mix at the end. You hear that kind of sound a lot now on female voices, like Christina Aguilera, and on a lot of really soulful R&B vocals. It's not there to fix the vocal; it's there to be part of the vocal sound.

"You can also use it to get a very cool portamento effect on vocals or on instruments," Binder continues. "When you get a nice R&B slide or slur in the vocal, Auto-Tune can enhance it and make it even smoother. I mean, it almost sounds calculated, like you can hear the algorithms processing as you do it. But that's become part of the vocal sound for a lot of singers now."

Binder has recently discovered a new toy: the rather expensive Waycom flat-screen monitor, which allows the user to write directly on its surface and affect the data on the screen, In the case of Auto-Tune, says Binder, this is the ultimate level of control. "You go into freehand mode and you can trace the graph and very, very accurately follow the wave of the vocal," he explains. "Once you get comfortable with the geometry, you can bend lines and smooth out all the imperfections. You can also get crazy with it — someone once drew a line that went off the graph and when we applied it to a guitar you could make it scream up three octaves and then dive bomb, all in under three seconds. It's a great effect for electronica and genres like that. It's going to change the world of Auto-Tune-ing."

But Binder, sounding wise enough for the age of 23, also opines that it's not the public that drives the use of auto-tuning, nor is it a lack of chops, in most cases. Rather, he says, it's frequently producers striving for perfection, often thinking of radio performances. "But whatever the reason someone uses it, it's not hard to tell that it's being used," he says. "For people who have heard a vocal track before its processed and then after, the difference is pretty apparent if you have reasonably decent ears. The new trick will be finding ways to use it and not have anyone notice."

Using Auto-Tune To Avoid Using Auto-Tune?

Joe Barresi, who has engineered for Queens of the Stone Age, Courtney Love and Hole, and Pennywise, has his own take on Auto-Tune. While he also tries to keep the application of the processing relatively light and close to the graph of the actual vocal, he has a surprising and somewhat ironic use for it. "It's actually a great tool to help singers with, and not by tuning their vocals," he says. "What I'll do, especially if it's a young singer or someone relatively inexperienced in the studio, is use the waveform graphic to point out where the problem areas or notes are. I use Auto-Tune as an analytical tool, giving singers a visual cue to show where the problem is. Very often, that's enough to get them on track and on pitch. Ironically, we're using Auto-Tune to avoid having to use Auto-Tune!"

Although the Antares Auto-Tune graphic display can be used to pitch-shift sections of a vocal manually, some engineers use it merely as a detection tool, either for providing visual cues for pitch processing in another application, or for providing feedback to novice vocalists in order to improve their tuning during performance.Although the Antares Auto-Tune graphic display can be used to pitch-shift sections of a vocal manually, some engineers use it merely as a detection tool, either for providing visual cues for pitch processing in another application, or for providing feedback to novice vocalists in order to improve their tuning during performance.

Barresi finds that Auto-Tune's inherent delay is usually easily compensated for by sliding the vocal track forward between two and four milliseconds. "It doesn't seem to be an issue for a lot of people, but I do like to hear the vocal stay in close sync with the track," he says. "So just applying a slide [globally] to the track from the first beat works well."

Less casual are Barresi's thoughts on how to pitch a vocal. "A lot of people do the tuning off-line, so to speak, as a stand-alone operation, as opposed to doing it while listening to the track," he explains. "I grab the vocal session graphically and load it into Auto-Tune and make the pitch adjustments while the rest of the track is playing. What I'm doing is listening for the pitch cues that the track can give you. For instance, you might have a vocal that's a bit sharp, but so's the guitar. It sounds worse if the vocal is in perfect pitch but the track isn't."

Another interesting approach Barresi has seen used is to work with Auto-Tune and Serato's Pitch N Time product simultaneously. "A lot of people prefer the way Pitch N Time sounds, though they like the way Auto-Tune works better," he says. "You can run both products together, using Auto-Tune to locate, analyse and determine the extent of the problem and Pitch N Time to correct it. For instance, if you see on the Auto-Tune screen that the pitch is off 25 cents, you can then create an anti-curve to the Auto-Tune graphic on Pitch N Time to correct it. Pitch N Time seems to affect the tone of the vocal less than Auto-Tune." (Although, Barresi notes, as do other engineers, that many young bands find the timbral effect of Auto-Tune to be a desirable tonal effect in and of itself.) Finally, he adds, Auto-Tune seems to track the vocal better if it's applied when the screen is zoomed out. "I'll work on a line or a phrase, then after I've corrected the pitch spots, but before I print it to a separate Auto-Tune track, I'll zoom it out first," he explains.

UK engineer Donal Hodgson, who has engineered for Sting, Counting Crows, Tina Turner, The Wilsons and Euphoria, has his Auto-Tune working preferences. "I regularly have Auto-Tune inserted in Pro Tools in automatic mode and will automate the bypass switch and the response time just to fix certain syllables or words," he explains. "I do the same on monophonic instruments as well. I have in the past had Auto-Tune on a monitor insert whilst recording a singer. I found this actually helped him with his ability to perform, as he stopped worrying about the small tuning problems and relaxed giving a much better overall performance. Again, when working like this I will slow the response time down so as not to induce the famed Cher vocal effect.

"In certain situations where a great performance has been recorded but bigger tuning problems have arisen, I'll switch to Manual mode, set up a record track and tweak lines manually. I always prefix the track name with 'AT' so that when I finish the comp I can see where all my tracks have come from. Manual mode is an amazing feature of Auto-Tune, and the artifacts of the tuning correction are virtually inaudible. Retuning like this with live vocal performances can be tricky, especially if the vocal has been picked up on other microphones — a certain amount of chorusing starts to happen and then a few subtle rides in the mix are needed to reduce the effect of retuning the vocal.

"I don't believe there should be any limitations on the resources used to reach a great-sounding vocal regardless of the singer's ability. I am hired to help make a recording that can be sold, and if the vocal performance isn't cutting it, then get the tool box out and fix it. Having said that, on the rare occasion I have been sent into the studio with someone who can't sing, all the Auto-Tune in the world isn't going to make them sound like a singer! I believe some talent is needed in the first place and then all the tricks can be added. I suppose this technology could be considered either as inducing apathy, or as a time saver — I have often fixed a vocal because it was quicker and easier than sending the singer back to the booth. I think it might be human nature."

Dry Is The New Wet

Meanwhile, the debate is not limited to DAWs. Effected vocals are a time-tested and legitimate artistic colour of the sonic palette, and have been undergoing an evolution quite apart from automatic pitch-correction. Some engineers will tell you that, while there are now more effects on vocals, the net result has been a drier sound than vocals effected with traditional reverb, mainly because most new effects track the vocal, rather than tail it as reverb does. (For instance, Auto-Tune can react in 4ms or less to the input signal.)

Engineer Michael Brauer: "Artists say they want drier-sounding vocals... but what they really mean is that they want something other than reverb."Engineer Michael Brauer: "Artists say they want drier-sounding vocals... but what they really mean is that they want something other than reverb.""Artists say they want drier-sounding vocals," says Michael Brauer, engineer for Coldplay, New Radicals and Sugar Ray. "But what they really mean is that they want something other than reverb. When you hear the vocal truly dry, it loses its life. So you go in search of other types of effects." Brauer's bag of tricks includes returning the vocal through various compressors, each of which has its own tonal characteristics. "That gives you a vocal that's still dry, but very present," he says. He has also sent vocals through guitar distortion boxes like the Sansamp, which when added in subtly gives an effect he describes as 'angry and urgent'. Bob Bullock has done similar things, and recalls that years ago he would use a Tom Scholz Rockman unit for vocals. Even on the country records he often works on, which tend to want to project an organic vocal sound, he has sent lead vocals through Leslie amplifiers and added chorusing.

Csaba Petocz, whose engineering work has ranged from Metallica to John Michael Montgomery, has used up to six Harmonizers at a time on a lead vocal, blending them in lightly to give a lead vocal more emphasis. He says the idea of heavily processing vocals has been around a long time, but that plug-ins offer a vastly wider array of sonic choices, all of which can be tried more readily. And therein lies the problem, he feels. "There are too many choices, which leads to mediocre records. To paraphrase Geoff Emerick, if you can limit your choices so that every decision supports a specific production goal, that's what makes great records."

Joe Chiccarelli, who has engineered for Beck, U2 and Elton John, says that the goal should be to support the song. "The easy availability of plug-ins too often leads to using effects for the sake of using them," he says. "Striking the balance is still the artistic part of it."

Published October 2003