Google Website Translator Gadget

Sunday, 1 January 2012

The most synthetic pop ever becomes an oddly democratic artform

If you're Japanese, you may have heard of Hatsune Miku. She(it?) is a Japanese pop star who(that?) doesn't exist. Back in 2004, Yamaha developed a signal processing application that allowed users to enter lyrics and a melody and have it rendered out as a singing voice. Stresses and intonation can be added an adjusted and other effects like vibrato and changing tones can also be applied.

It was initially used by musicians to add backing vocals to tracks without having to get a performer into the studio. People like Mike Oldfield started using it and Yamaha started producing singers-in-a-box in association with various companies - you bought one software package containing Vocaloid and a single singer's voice. It went pretty and soon there were a range of voices available.

Vocaloid 2 interface

When Yamaha released the second version of the Vocaloid software, Crypton Future Media (who had previously developed the Meiko and Kaito voices) created Hatsune Miku and everything went nuts. Hatsune Miku was based on the voice of the very real anime voice actress Saki Fujita who had also sung the theme tunes to a few anime series. For some reason, the Hatsune Miku voice became massively popular and soon hordes of Japanese were banging out Hatsune Miku songs in their bedrooms and studios.

By August of 2010, just three years after the release of the HatsuneMiku Vocaloid, it was estimated that there had been over 22,000 original songs written for it and over 100,000 sings including cover versions. There was just one problem: Hatsune Miku was just a voice and although the pop-buying public will buy music without ever seeing the performer, there was a real desire to have a visual Hatsune Miku.

MikuMikuDance software
Yu Higuchi released the freeware MikuMikuDance which allowed users to pose and animate characters to create videos to go along with their Hatsune Miku songs. It soon became very popular and hundreds of creators have released new characters for use in the program; which itself is constantly updated with new features.

By this time, Vocaloid albums were being released and charting well. One album (Exit Tunes Presents Vocalogenesis feat. Hatsune Miku) opened at the number one spot on the Japanese Oricon charts and went on to sell very well indeed.

By now the only thing missing from the Hatsune Miku phenomenon were live concerts. But thanks to digital projection on transparent screens, Hatsune Miku performed her first, ahem... live concert in August of 2009.

It might seem pretty crazy that people actually paid to go and watch a concert given by a synthetic voice and a projected rendered image, but it was a huge success. A series of transparent screens and eight projectors made for an almost 3D holographic style of projection. What I find particularly odd is that the band plays live.

Now just in case you're thinking this is some mental thing that only Japanese people get involved in, there have been concerts performed worldwide. There's even one coming up in Scotland in which you can get involved if you'd like. And that's what's become so very interesting about this. We're all used to synthetic music these days; whether it's the Simon Cowell-manufactured acts that dominate the UK pop charts or the sample-heavy DJ-led acts that make up the club scene. What's different about Hatsune Miku is that she doesn't really belong to anyone and so, in a very real sense, she belongs to everyone.


  1. Vocaloids are alright for J-Pop and other squeaky clean female vocals but they sound worse and worse the lower down the range they get. To get any decent results also needs a lot of tweaking, so most of the end product is total shite. I had considered getting one but I'm not that impressed.

    I'm not sure about the democratisation aspect. If the results of other music production automation is anything to go by, it will just mean a load of amateur chod, a plethora of niche semi-pro stuff that makes money mainly for the distribution infrastructure providers, and a top end dominated by empty marketing constructs aimed at ignorant, faddish 14 year olds. Plus ca change....

  2. Some very fair points. In film-making, digital cameras have lowered the bar to entry, but conversely whilst they've made it easier to make a film, I think they've made it harder to make a good one. When films were all shot on 35mm film stock, the sheer costs involved tended to weed out the crap in a sort of self-selecting process. If you can't convince a dozen people that your idea's a good one just to get the funding to roll the cameras, then maybe your idea isn't that good.

    So by making it easier to make music and the visuals surrounding it, I reckon it's a similar process that will indeed open the floodgates of crapitude. But it might just allow for some good experimental stuff that would never have got a record deal otherwise. Hmm...

    Inevitably, Vocaloids will improve in quality and you're quite right that they're best suited to squeaky jpop at the moment. But they're already appearing in mainstream pop as backing singers without anyone noticing so it'll be interesting to see if, in another couple of generations' time, whether one will succeed in the western pop world. There have already been a couple of projects to resurrect dead singers' voices through Vocaloid technology and those have been interesting if a little creepy and exploitative.

  3. Also creepy are the Vocaloid "community", who seem obsessed with imagining the robovocalists to be real people and trying to work out which of them are having sex with each other... (shudder)

  4. Oh, too frigging right. When they start working out blood types and and arguing about the romantic compatibilities between Vocaloids based on them, that just freaks me out. How long's it going to be before they start making sex dolls that look like their favourite Vocaloids? That's if they haven't already... (shudder)

    There are already Japanese sex comics (hentai doujin) about Vocaloids. Eep!