This isn’t a joke: What do Cher and ExxonMobil have in common? The answer is mathematician Dr. Andy Hildebrand, the inventor of Auto-Tune. Cher‘s iconic “Believe,” which popularized the warbly, “robotic” sound long before T-Pain or Kanye West did, turns 20, so we’re taking a look back at the technology that’s helped — or hindered — pop music ever since.
Despite being a professional mathematician of sorts, Hildebrand wasn’t a musical dilettante: A classical flute player, he was a studio musician by the time he was 16 and paid his way through college, in part, by teaching flute lessons. Years later, he was at a National Association of Music Merchants (NAMM) conference where, during lunch, a woman made a comment that what the music world really needed was something to help her sing in tune.
To avoid a very long explanation that weaves into seismology, acoustics and math, let it just be said that Hildebrand basically figured out he could invent such a device using the same technology that he’d been working with for Exxon.
Prior to Auto-Tune, it was common practice for a singer to have their final take on a recording pieced together from multiple attempts. Before digital recording, this involved slicing up the actual tape, an undertaking magnified or lessened by the singer’s pitch accuracy. Hildebrand was seeking a way to marry the “live” aspect of recording with the scientific advances he’d wrought: “The singer’s first take is often their best, it’s full of vitality and emotion,” he told NPR in 2004. “After the take, their producer will announce ‘Great, but the second phrase was pitchy so let’s do it again.’ Well, now the singer’s worried about pitch and has to focus on the intonation and the vitality and emotion are gone from their performance. What Auto-Tune lets the producer do is fix the first take.”
“In my mind it’s not very complex,” he’s said of the theory behind his invention, “but I haven’t yet found anyone I can explain it to who understands it. I usually just say, ‘It’s magic.'”
For Auto-Tune to work correctly, the user sets the song’s key — so it can “know” how to adjust pitches — as well as rate at which the program alters wonky notes. Hildebrand built Auto-Tune with a sliding scale to govern that speed, from 1-10. “Just for kicks,” he said, “I put a ‘zero’ setting, which changed the pitch the exact moment [the program] received the signal [the recorded voice/keyboard/what have you].”
This is — more or less — where Cher comes in.
Songwriters Brian Higgins, Matt Gray, Stuart McLennen and Tim Powell had been shopping “Believe” around Warner Bros. for months to no avail. “[Executive producer and former chairman of Warner Brothers Rob Dickins] asked us if we could sort it out,” producer Mark Taylor told Sound on Sound in 1999. “Two of our writers, Steve Torch and Paul Barry, got involved and eventually came up with a complete song that Rob and Cher were happy with.” Dickins’ idea was to give Cher a dance single that would — presumably — capitalize on the enormous success Madonna had found with Ray of Light, released in February 1998. “The hard part was trying to make one that wouldn’t alienate Cher’s existing fans,” Taylor explained. We couldn’t afford to have anyone say ‘I hate this because it’s dance’ — then we would have turned off loads of people who are used to hearing Cher do rock ballads and MOR songs.”
Cher — who, contrary to the song’s futuristic bent, recorded her vocal partially with a Neumann U67, a very expensive, vintage tube microphone she’d just used on a session with George Martin — does not have a credit on the song as a writer, though she’s since said she altered the demo’s lyrics. “I was singing [the song] in the bathtub, and it seemed to me the second verse was too whiny. It kind of pissed me off, so I changed it. I toughened it up a bit. I wrote the lyrics, ‘It takes time to move on, it takes love to be strong / I’ve had time to think it through and maybe I’m too good for you.'”
During production, Cher happened to luck into a solution for juicing what the team all agreed were the song’s lifeless verses. She’d been watching TV and saw a singer named Andrew Roachford, whose album she then purchased. ”We were tackling ‘Believe’ for the gazillionth time,” she told The New York Times in 1999. ”And I said: ‘I’m so tired of doing this. Let’s just put on this CD and listen to music and get away from this.”’ One of the songs featured Roachford’s vocals manipulated by a vocoder, another kind of sound synthesis device adapted for music by forward-thinking individuals like Wendy Carlos, Robert Moog and Phil Collins. Cher suggested they try the same trick. Taylor — who, out of fear of exposing industry secrets, muddied the “Believe” water by insisting the effect was done with a vocoder — applied the “0” setting on the Auto-Tune software and struck gold, even if he was too afraid to share his work at first. “A couple of beers later,” he told the Times, “we decided to play it for her, and she just freaked out.”
Cher concurs: “We high-fived,” she told the Times. ”It was like some stupid Rocky film.” [Editor’s Note: Cher is the best.]
Not everyone agreed, and the singer had to throw some weight around to get the track released with the novel effect. Dickins, who’d originally been pushing for a dance sound, thought the group had gone too far. ”He said, ‘Everyone loves that song but wants to change that part of it,”’ Cher said. ”I said, ‘You can change that part of it, over my dead body!’ And that was the end of the discussion. I said to Mark before I left, ‘Don’t let anyone touch this track, or I’m going to rip your throat out.”’ [Editor’s Note: Cher is the best.]
Now, just to refresh everyone’s memory, this is how “Believe” performed when it hit the airwaves: It peaked at number one in 23 countries, it was in the top 40 of the Billboard hot 100, it solidified Cher as the oldest female artist (52) to achieve all of this.
And, then, obviously, came the backlash. South Park went with a bit suggesting that “Believe” could be used as psychological torture…
… while Buffy the Vampire Slayer took a slightly different route, suggesting that a literal college roommate from Hell would be playing the song on repeat (while griping about her missing milk).
Auto-Tune would go on to enjoy a checkered reputation, becoming lazy shorthand for the argument that technology was running rampant on music. Then, of course, there was T-Pain. Between 2005 and 2009, three of the former Faheem Rashad Najm’s heavily Auto-Tuned records went platinum; during one landmark week in 2007, he was featured on four different Top 10 Billboard Hot 100 singles simultaneously. As was made clear by his Tiny Desk concert and his own admission, T-Pain can sing, he just thought Auto-Tune would set him apart from the pack.
Since The Reign of T-Pain, and the subsequent backlash to and rem-embracing of egregious pitch correction, you can see descendants of “Believe” all over if you look hard enough. Kanye West (who apparently thanked Cher for popularizing it when they met in 2015) used it extensively on 2008’s 808s & Heartbreak and various onstage rants since …
… while Future has used the effect to give himself a disaffected quaver that suggests the alienating effects of, well, you know… all the drugs.
Bon Iver, following the rudimentary recording techniques of For Emma, Forever Ago, began exploring Auto-Tune and now uses a much more sophisticated version of voice manipulation, developed by his engineer, Chris Messina, for the eerie choral effect that permeated his most recent record, 2016’s 22, A Million.
Cher, for her part, remains a careful steward of her work: You can find her defending her and the rest of the team behind “Believe”‘s innovation on Twitter.
Don’t worry, Cher. To paraphrase one of your peers in meme-dom, we haven’t, and won’t ever stop believin’.