It’s been a strange journey for the robot voice. The usual trajectory for new bits of musical technology is to begin somewhere out on the leftfield, gradually growing in popularity until they choke up the mainstream, before quickly dying out. Take scratching: pioneered as a new sound by a small group of New York DJs in the early ’80s, it became progressively more widely accepted until, by the late ’90s, it was being so heavily used as an easy marker of -urban’ credibility – appearing routinely on nu-metal tracks, skateboard video games, and even Hanson’s puberty-rock anthem -Mmm Bop’ – that it plummeted, deservedly, out of favour.

Auto-Tune didn’t start out on the margins. It began on a single that sold upwards of 10 million copies, sung by an ageing gay icon, chiefly known for revealing outfits and a strange grimace brought on by considerable plastic surgery. It then flirted briefly with credibility on Daft Punk’s -One More Time’, before being endlessly imitated until it appeared only on the kind of production-line Eurohouse that people play out of their mobile phones at the backs of buses. Now, suddenly, it’s all over us again.

T-Pain, who has made the sound his own, has released two of the biggest selling albums in the States these last two years. Kanye has plastered it all over his new release. And Lil’ Wayne – the critics’ darling – used it extensively on one of the most highly-anticipated albums of 2008, Tha Carter III. How can this be? You might put it down to 2007′s revival of interest in European dance stylings. The crossover appeal of Justice and friends, along with renewed success for Daft Punk, brought the Robot Voice back into the fold and lent it a leftfield credibility it had never previously enjoyed.

However, Nic Bertino suggests that there might be a more definite musical reason why people have latched onto the Auto-Tune effect: ‘It brings the vocoder sound, which is very mechanical, and mixes it with a human sound – so it’s a human sound that’s just a little bit robotic.’ Vocoders are an earlier and more primitive way of electronic-ising the voice (think of the hook on Tupac’s -California Love’). With a vocoder, you get the robot, but it takes the human out. With Auto-Tune, you have the electronic sound, but you don’t lose the capacity for the singer to express emotion.

Oddly enough, expressing emotion seems to be exactly what Auto-Tune gets used for a lot of the time. Cher’s -Believe’, for all its core of frothy dance-pop, is a breaking-up song; and Auto-Tune’s glitching occurs most on the lyrics where feeling runs highest: ‘I can’t break through’¦ So sad that you’re leaving.’ 808s And Heartbreak was recorded after the death of Kanye’s mother and the breakdown of his relationship with his fiancée.

On -Love Lockdown’, there’s no rapping at all. Yeezy has lost his Louis Vuitton Don swagger entirely, and in its place is a resigned-sounding ode to a failing relationship (‘You lose, you lose’). It’s almost as if singing was the only way to express how he feels. ‘It was just what was in my heart,’ he has said of his singing. ‘The melodies were in me.’ Elsewhere, he’s described the album as ‘a little bit of Auto-Tune and a whole lot of fucked-up life.’ This combination, however, is exactly what Kanye’s detractors on the blog seem to be complaining about. One summed up the general feeling: ‘This is one of the most personal of his songs. I want to hear Kanye, not a computer.’ People will accept practically any kind of electronic distortion or synthesised sound on a track. But when it comes to the human voice, there’s this idea that it’s somehow -truer’, the more raw it sounds. And when Kanye alters it electronically, people act like he’s covering something up. ‘PLEASE re-do this’, wrote another commenter, ‘and SING YOUR PAIN OUT!’

In 1967, when Marvin Gaye was recording -I Heard It Through The Grapevine’, his producer arranged the track in a key slightly higher than Gaye’s vocal register, so he’d have to strain to hit the top notes. The point was to get a rawer sound, as though Gaye’s voice was cracking with emotion. The trick was a success – but it’s a trick. Gaye was singing words written by and for someone else and his emotion is no more genuine than if it had been added with a computer. So why should -I Heard It Through The Grapevine’, a real vocal about a made-up heartbreak, sound any more -true’ than -Love Lockdown’, which deals with real heartbreak through a made-up vocal? The bottom line, it seems, is that when it comes to recorded displays of emotion, we’re happier with a convincing fake than we are with a fake-sounding reality. But perhaps that’s no bad thing. The point of a song, after all, is to soundtrack our own experiences, not the artist’s. That’s what music is there for. So in a way, it doesn’t matter how genuine Kanye’s heartbreak is: if we can’t translate it into our own, it’s never going to have soul.