Bobby Miller
Writer Website
Sing, Computer, Sing: Vocaloid Technology
1,300 words
Ever since Cher auto-tuned her song “Believe” in 1998, the line between an actual human voice and a digital one has slowly blurred. However, auto-tuning still requires a human singing. But thanks to technological advances by the music company Yamaha, voices can now be produced with just a computer program.
This program is called Vocaloid, a combination of vocal and android. According to Vocaloid.com, “the software allows users to input melody and lyrics in order to synthesize singing. In other words, with this technology, singing can be produced without a singer.” Everyone who uses the program has a database of sounds at their disposal, the “phonemes” that make up language. By combining these language sounds at will, they can produce song lyrics.
To produce the voices used by Vocaloid, a company must record a human voice actor making all the possible sounds in a given language, which is no small task. However, once this database is created, the program users can manipulate the voices as they please. According to Yahama’s Vocaloid FAQ, the user is allowed to “create vocals for commercial or non-commercial use as long as the vocals do not offend public policy” with “derogatory or disturbing lyrics.” So, much as I can use Microsoft Word to type up something I own, people can use Vocaloid to compose something they own. Users can sell their work on websites like iTunes or Amazon, though many post them on YouTube or the Japanese video website Nico Nico Douga.
What do the composed songs sound like? Well, for right now, nobody will be mistaking the computer-generated voices for authentically human ones: they have a “robotic” sound to them. This does not bother Vocaloid users, however. In a discussion with Wired.com, Yahama researcher Hideki Kenmochi said, “Especially in Japan, Vocaloid is not regarded as a substitute for human singing, but a kind of new musical instrument.” Enthusiasts are free to use this instrument as they please, commercially or just for fun.
The only catch is that users are not allowed to use the Vocaloid character mascots for commercial purposes without permission. You see, every Vocaloid voice has a character associated with it. The most famous of these characters is Hatsune Miku, whose appearance and voice were both created by Crypton Future Media, Inc. The company specializes in making music products from new technology, as its website explains. Although Yahama made the Vocaloid software, other companies are indeed free to make their own voice databanks if they wish.
Crypton used samples from Japanese voice actress Saki Fujita to create Hatsune Miku’s voice. Since her release on August 31, 2007, the character’s voice and anime-like appearance have won her many fans. She is currently the most popular Vocaloid voice out of the dozens available.
Vocaloid’s Popularity
In fact, according to ClashMusic.com, Hatsune Miku is actually “Japan’s biggest selling pop star”—quite an accomplishment for someone who doesn’t really exist. As reported by the Anime News Network, compilations of Vocaloid music have sometimes topped CD selling charts, beating out all the “real” singers.
Hatsune Miku’s concerts have also been successful. That’s right: this computer program even performs live. As the song is played, a hologram version of Hatsune Miku appears onstage, singing and dancing to the music. In some concerts, she does not appear as a hologram; instead, a large screen shows her dancing.
Vocaloid has even appeared in a rhythm game of its own called Hatsune Miku: Project DIVA. Developed by Sega, this PlayStation Portable game has players pressing buttons in time with popular Vocaloid songs from various characters. Since its release in 2009, the game has received four sequels. The most recent one, Hatsune Miku: Project DIVA F, sold over 150,000 copies during its first week available, according to 4gamer.net. The runner-up that week sold only a third as many copies.
While the official, large-scale projects are exciting, the fans are what keep Vocaloid alive and well. Thousands upon thousands of fan-made videos can be found online. For particularly ambitious users, there is a fan convention called The Voc@loid M@ster held four times a year in or near Tokyo, according to the Miku Database website. Here, fans can sell any albums they have composed or share any other artwork pertaining to Vocaloid.
Yes, Vocaloid’s artistic appeal lies not just in music but also in visuals. The anime-style characters behind each voice have resulted in all sorts of fan art and homemade music videos. Fan art of Hatsune Miku is so popular, in fact, that searching for her on Google Images generates 23 million results. To that in perspective, keep in mind that Sonic the Hedgehog, a video game character known worldwide, generates only 16 million results. Aside from Hatsune Miku, other famous Vocaloid characters include Kagamine Rin and Kagamine Len, a blond girl and boy, respectively.
Fans have interpreted each Vocaloid character in different ways. Some artists try to depict them realistically, while others opt for a more cartoony style. There is also a great deal of erotic art based on the characters.
But Japan’s top-selling pop star is almost unknown in the United States. Sure, there have been Hatsune Miku concerts on our shores, but they usually take place in festivals devoted to anime and Japanese pop music. In other words, she only appears in places where audiences have a love for Japanese culture. Although the program is available in both Japanese and English, Vocaloid is an obscure technology to most Americans.
Vocaloid’s popularity is on the rise, however. About half of Vocaloid downloads from iTunes come not from Japan, but from various countries worldwide. This statistic from the blog of Piapro, a Crypton website dedicated to Vocaloid uploads, shows that interest for the program does exist beyond Japan.
Recent Developments
The most recent version of the program, Vocaloid 3, should help to broaden the appeal of this technology. According to Vocaloid.com, the text in this release can be read in English, Japanese, Chinese, Korean, or Spanish. However, the actual voices are still limited mostly to Japanese and English.
New languages aren’t the only additions made to Vocaloid 3, however. The program allows the user to record their own voice and have it converted into the Vocaloid voice. So, a user can sing a song and then hear Hatsune Miku sing it back to them. This makes the program easier to use for people who are not programming experts. Researchers Tomoyasu Nakano and Masataka Goto have posted demonstrations of this technology on YouTube and the official website for the National Institute of Advanced Industrial Science and Technology (AIST). Yahama could not be reached to comment on how these additions might broaden the appeal of Vocaloid.
The possibilities are exciting. By broadening the number of languages that the program supports, it should slowly find a larger audience beyond Japan’s borders. Plus, Vocaloid could be used in a number of different ways. As the New York Times has speculated, this voice-synthesizing technology might one day be used to bring dead singers back to life. By collecting various parts of Elvis recordings, for example, one could build a phoneme database of his voice and use it in any way imaginable. According to Wired.com, some researchers are trying to bring back the Japanese singer Hitoshi Ueki by using voice clips of his singing.
The realism of Vocaloid’s singing will grow with time, making the line between “real” singing and synthesized singing even blurrier. The technology can allow any person to generate a quality song. The possibilities in store for the program will continue to grow as it reaches a broader audience worldwide.
Bobby Miller originally wrote this article for the digital magazine Nvate in October of 2012. Due to its age, some of the information or links in the article may be outdated.