Score one for creepy technology. Vocaloid, a voice-synthesis brand owned by Yamaha, has come up with a process by which to “resurrect” any singer’s voice for use in synthesized songs, without requiring the vocalist to build up a painstaking voice library first – so they could be doing that Kurt Cobain/Michael Jackson duet album pretty soon.

Typically the process of synthesizing an artist’s voice would require them to sing every single possible syllable in the target language, one at a time, which would later be stitched together to form songs. This takes time, effort, and requires the singers to be alive to do it.

Vocaloid claims, however, that they’ve succeeded in bypassing the recording phase, as displayed by their resurrection of Hitoshi Ueki, a popular Japanese vocalist who died back in 2007.

Said Yamaha researcher Hideki Kenmochi in an interview with Wired,

As far as I know, many viewers were satisfied with the result, and so am I. It really sounds like him, because the creator [the programmer in charge of the voice library] did a good job.

The technology isn’t perfect by any means – the uncanny valley effect that makes overly-detailed CGI faces look creepy also applies to synthesized music, and there are times when it’s clear that a bit of voice was computer-generated.

Still, the methods Vocaloid are employing are so promising, in fact, that they’re planning to release an album with the robot Ueki-voice in a year or so.

[Source: Wired]