Join 3,557 readers in helping fund MetaFilter (Hide)


I suppose my voice will always fall short
January 3, 2014 2:06 PM   Subscribe

Yukkedoluce is a producer for vocaloid, a singing voice synthesizer program (previously on mefi). Some examples of his works are below.

Transient Apple Salesgirl's description is from vgperson's subtitled video, linked above. You can find art/video credits for Transient Apple Salesgirl here on Nico. Yukkedoluce himself did the art and video for the latter three.

Yukkedoluce has his own Youtube channel, which he updates pretty regularly. You can also check out his works at Nico Nico Douga. He has released one album, which I don't believe is available on the Western Amazon sites or iTunes...but you can still import it from Japan. (There is one song available on both Amazon and iTunes, though.)

He utilizes several vocaloid programs. The above songs, for instance, are sung by GUMI (or Megpoid), Hatsune Miku, and Kagamine Rin.
posted by anthy (15 comments total) 11 users marked this as a favorite

 
And, of course, vocaloid was used for the voice of Puppycat in Bee and Puppycat.
posted by heathkit at 2:24 PM on January 3


Can it do singers who aren't schoolgirls or other anime characters?
posted by thelonius at 2:29 PM on January 3


Vocaloid previously on MeFi Music.
posted by teraflop at 3:06 PM on January 3


thelonius: there's a number of vocaloids that have been released without a character concept. On the page you link to you can find VY1 and VY2, a case of "character intentionelly left blank", intended to encourage the community's creativity and to allow them to create their own avatar to go with the music.

But there's also Prima, a soprano opera vocaloid. And Lola, which was used for the original soundtrack to Satoshi Kon's Paprika. And there's probably a bunch more that I don't even know about.
posted by Caconym at 4:09 PM on January 3 [1 favorite]


thelonius, nobody's ever tried.


There is a game called Hatsune Miku Project Diva F on the PS3 or PS Vita that is utterly consuming me right now. If you enjoy rhythm games (which are truly in short supply outside of Japan) I urge you to get it. The game draws up some of the best Vocaloid tracks of the last few years and pairs them with intense button-tapping action and extremely bright animations (which, once you've beat the level, you can watch as videos). It also has dress-up games and some kind of Tamagotchi mode.

This song/animation NegaPosi*Continues is simply wonderful, and its intense rhythmic changes make for a challenging level to beat.

Another great one is Online Game Addicts Sprechchor which has lyrics about turning away from the real world to play MMO games for the rest of your life. But it's upbeat!
posted by sixohsix at 4:12 PM on January 3


...okay Caconym corrected me. I guess there are Vocaloids without anime heads. But not usually for very long.
posted by sixohsix at 4:13 PM on January 3 [1 favorite]


It sounds surprisingly good
posted by thelonius at 4:18 PM on January 3


Can it do singers who aren't schoolgirls or other anime characters?

...and does anyone use Vocaloid to create music other than anime-flavored bubblegum J-pop? Cuz, uh, I'm not a fan of that stuff. But it's an interesting tool, and I wonder what else it might be capable of.
posted by escape from the potato planet at 7:19 PM on January 3


That opera vocaloid is remarkable. If one wanted to create an original composition using such a vocaloid, where in the world would one start?
posted by stolyarova at 7:59 PM on January 3


Have to say I wasn't aware of Prima! It can actually render english pretty well!

I've always been disappointed in vocaloids until now as I wasn't aware they could manage anything other than bubblegum J-pop junk. Granted, some of that stuff is still really amazing. I wasn't even aware they'd tried to make a masculine vocaloid until a friend got hold of one of the oddball PSP games.

I think part of the problem with the more advanced stuff catching on is that the idea of Hatsune Miku not as a tool but as a character/personality has taken off more than anything else and so it gets the majority of attention. I'd still how William Gibson feels about the whole thing after having written Idoru.
posted by ThrowbackDave at 2:01 AM on January 4


and does anyone use Vocaloid to create music other than anime-flavored bubblegum J-pop?

Something I found while searching for random shamisen covers: what happens when you take a retro-look vocaloid song like Senbonzakura ("Ten thousand cherry trees") and play it with koto, shamisen, flutes and drums is this awesomeness.
posted by sukeban at 2:02 AM on January 4 [2 favorites]


sukeban, that was pretty great. Thanks!
posted by escape from the potato planet at 8:52 AM on January 4


> Have to say I wasn't aware of Prima! It can actually render english pretty well!

Most Vocaloids are build on Japanese morphemes but there are some designed for English. The first voices (Leon, Lola and Miriam) were English speakers. There are also Spanish, Korean, Chinese, and one Catalan.

For a while now I've entertained the idea of doing a cover of Radiohead's "Fitter, Happier" with Hatsune Miku instead of that text-to-speech computer voice, specifically because it wouldn't necessarily map English accurately. Learning new software and going through with it felt like too much effort for a one-off joke, though.
posted by ardgedee at 9:41 AM on January 4


ardgedee, I think you should do it. Incidentally, Hatsune Miku V3 is available with an English voice, which would make it easier. 150 USD (also includes the vocaloid editor and a whole bunch of other stuff I don't know much about).
posted by sixohsix at 11:27 AM on January 4


For this particular thing, it would specifically have to be a Japanese-language Vocaloid, whether or not it's Miku... The original is relayed through Macintalk in part because doing so makes its horrific shopping list of affirmations and goals harder to pin down emotionally or to identify with anybody. There is pointedly no human reading those words.

Seventeen years later, we've got Siri and her e-kin from Google and Microsoft fulfilling the technological promises of ultimate personal organization offered by the Newton and Palm Pilot and text-to-speech. The computer is just as impersonal even though it's doing a much better job of faking a connection. It would be trivial to record Siri reading a transcription back but that's not what I had in mind. Siri's urgent but unhassled tone represents the present, and it also has a cultural ubiquity that Macintalk never quite achieved. Also, having Siri read off the lyrics feels like working for a cheap and easy punchline directed at Siri specifically when something bigger could have been made.

Robot voices of the future will have to parlay a more complex emotional character than the current state of the art can achieve. If a synthetic voice is used for the purpose of depersonalizing and genericizing the alienated person portrayed in "Fitter, Happier", the voice has to be inappropriate to the content in some way. Instead of being monotonal, reedy, arrhythmic yet mechanical-sounding, what if it sound like it could barely keep from singing? What if it have so much difficulty pronouncing even the trailing R in "fitter" but moving along unperturbed by its own errors, and by the substance of what it's saying?
posted by ardgedee at 1:11 PM on January 4 [2 favorites]


« Older The town of Rjukan, Norway (Google Maps) lies in a...  |  Rory and Paris: The Real Gilmo... Newer »


This thread has been archived and is closed to new comments