VOCALOID3: New Features and Voicebanks (Updated)

The 'i-style project' by CARNELIAN

Vocaloidism Exclusive Info: you’ve all been waiting for it, VOCALOID3 is coming this year at the end of September! The Secret Party held on June 8 is over, and they’ve finally announced VOCALOID3, and then given us a lot more delicious information on all things Vocaloid.

… wait, you already know about that? Then, are you sure you know about YAMAHA’s special plans for building an army of VocaloP’s, the bright future of the Vocaloid culture, the range of price for VOCALOID3 Editor, the VocaListener Job Plugin, and the survey about the Vocaloid fandom? Hm… you know that too? Well then, one thing’s for sure: you don’t know what I ate yesterday, so read on for more Exclusive Info!

Update: demo song of the Korean Vocaloid voiced by Kim Tahi.

First, if you’re wondering why it took so long… I managed to write a very messy article just before going on a trip, but I just didn’t have the time to add the finishing touch — as in, it really couldn’t be published as is. Well, I’m back, and here’s the article five days later. ^^; Thanks to Krystal for helping getting it done!

Anyway, back to VOCALOID3; hopefully you’ll learn a few bits of additional info. Some of it may be outdated or incorrect, so feel free to let us know in the comment section!


At the secret party there were video messages from many company representatives. From Crypton was Hiroyuki Itoh, representing Zero-G was Dom Keefe, and for Power FX was Bil Bryant. Many other people important to the Vocaloid scene were also present at the event, including Internet’s Noboru Murakami, VocaloRevolution’s HirotoP, Ring Suzune’s voice provider MiKA, and famous anime singer Lia of Tori no Uta fame.

Features of VOCALOID3:

The new VOCALOID3 Interface. From: Gigazine

Here is a list of the features of VOCALOID3:

  • It will have better quality audio: longer samples will be used (probably similar to UTAU VCV voicebanks), sibilants such as /s/ will be less “metallic-sounding”, better /w/ sounds, smoother pronunciation including for fast-talking. The transition between different voice registers (low/high) will be smoother.
  • It will be easier to use: there is an improved interface, with a new display of the different tracks similar to what can be seen on DAW‘s, audio files (for off-vocal tracks for instance) can now be imported, and it will now work as a VST host (it will be possible to add reverb and other effects).
  • It will be possible to undo or redo any previous action in the Editor.
  • It will be compatible with VOCALOID2 voicebanks. However, it won’t be compatible with every one of them, and there may be a price to import some of them.
  • It will support these new languages: Korean, Chinese and Spanish (pinyin input for Chinese and hangul input for Korean possible). The Editor will also be localized.
  • It will work on Windows XP/Vista/7.

Although VOCALOID3 only works on Windows, Kenmochi said they want to release a Mac version “as soon as possible” because it’s “very important to [them]”. He also said that they could confirm it would work well, since they have already developed VocaloWitter (previously iVOCALOID-VY1t), which runs on the iPad and iPhone’s iOS (which is derived from Mac OS X).

VOCALOID 3 also features the VOCALOID Job Plugin. This is a function that will allow exterior plugins to control the Editor’s parameters and notes. At the moment YAMAHA is developing a VOCALOID Job Plugin version of Vocalistener, and will most likely sell it separately. They’re also considering making it compatible with English lyrics input, but they’re not sure they will make it on time. The price for Vocalistener hasn’t been announced yet.

The API of VOCALOID3 will be made public to licensed companies, whilst the technical details about the VOCALOID Job Plugin feature will be made public so that users can develop their own plugins compatible with VOCALOID3 and sell them on the VOCALOID STORE. YAMAHA will also start granting licences for the synthesis engine itself so that it can be used in things such as PC games for instance.

However, this awesome new VOCALOID3 Editor will actually be sold separately from the voicebanks, contrary to the previous versions. It will be sold on VOCALOID STORE; there is currently no price for it editor yet, but they think it will be around 10,000 yen (125 USD). A trial version of the Editor with limited features, called the “Tiny VOCALOID3 Editor”, will instead be shipped with the different voicebanks.

Here are the differences between the different Editors (VOCALOID3/Tiny VOCALOID3):

  • Number of tracks: 16/1.
  • Maximum number of bars: 999/17.
  • Effects: any effect available through the VTS Host/Reverb.
  • VOCALOID Job Plugin: Yes/No.
  • VOCALOID2 Library Import Tool: Yes/No.

New and old voices for VOCALOID3:

Eri Kitamura will be voicing CUL

With the VOCALOID3 engine coming up, the many companies who are currently releasing voicebanks have begun making plans on how to use the new engine. Here is what you can expect from your favourite companies.

  • AH Software:
    • AH Software is planning on making VB’s based on many singers. They also want to use Vocaloid to produce “different types of sounds”, not limiting them to only simulating singing. AHS plans to make Vocaloid “evolve”.
  • Crypton Future Media:
    • Crypton has enjoyed success so far with their current VB, and are planning on keeping their existing VB’s on VOCALOID2. There are no plans to re-release them on V3, but converting them to V3 will be free.
  • Internet Co., Ltd:
    • Internet Co., Ltd will be releasing a VB for VocaloRevolution’s CUL, who was voiced by VY1 until now. She will be voiced by famous seiyuu Eri Kitamura.
    • Gakupo, Gumi, Lily and Gachapoid will have a V3 update.
    • Internet is also releasing Megpoid Extend (tentative). Megpoid will have added four different types of voices: Power, Whisper, Adult and Sweet. It will be released on VOCALOID2, but will also have a VOCALOID3 version.
  • Power FX:
    • For VOCALOID3, they plan to release one new English-speaking Vocaloid.
  • YAMAHA:
    • The VY series will be updating to VOCALOID3: Yamaha will sell VY1 for V3 and VY2 for V3. Both of them will be sold together with the full Editor.
    • YAMAHA will also be releasing a new voicebank based on Miu Sakamoto (坂本美雨, the daughter of legendary electronic music band Yellow Magic Orchestra‘s Ryuuichi Sakamoto).
  • Zero-G:
    • Zero-G has plans to release two new English-speaking Vocaloids.

The Korean company SBS Artech will release a Korean voicebank; the voice provider will be artist Kim Tahi (17 year old, part of the K-pop girl band “Glam”). She will have both a Japanese and a Korean voicebank, and English pronunciation will be possible by using the Korean one. The character will be based on Tahi (or “Dahi” depending on the romanization system used), and will be “light-hearted and cute”. The voicebank is planned to be released this Fall.

Here’s the sample song, “I=Fantasy”, with Korean and English subtitles:

You can also watch it here as presented by SBS Artech during an earlier press conference.

The company 1st PLACE will be releasing its first Vocaloid, based on the singer Lia.

Ring Suzune and Lui Hibiki will also be released on VOCALOID3.

The Future of Vocaloid Culture:

MOEJAPAN has plans!

YAMAHA conducted a street survey with males and females from age 14 to 50; it was conducted in late April 2011 in different districts of Tokyo. According to the survey, 68.4% of peoples up to 20 years old, 61.1% in their twenties, 41.5% in their thirties, and 35.2% in their forties know about Vocaloid. About 1/4 of the girls in their teens often listen to Vocaloid music; 8.4% of them only listen to Vocaloid music, and 15.6% of them listen to it a lot (it constitutes more than 30% of what they listen to).

General statistics about VOCALOID and computer music (called “desktop music” or DTM for short in wasei-eigo) in Japan as estimated by YAMAHA are:

  • 5 million casual Vocaloid listeners
  • 100,000 core Vocaloid listeners
  • 10,000 Vocaloid producers (100,000 DTM [non-Vocaloid] producers, 10,000 on a professional level)
  • 8,000 creators of derivative works (manga, cosplay, goodies)

YAMAHA wants to promote the Vocaloid culture and help it grow by:

  • helping aspiring P’s by providing “P training” classes
  • creating a record label, VOCALOID MUSIC PUBLISHING, that will also deal with copyrights among other things and will be “flexible” to match the users’ need, and will serve as an intermediate with JASRAC for instance for partial registration. It will also help doujin circles to exchange information
  • encouraging the creation of new characters (through the Everyone’s Vocaloid Project for instance)
  • being more present in the entertainment world (e.g concerts, games)
  • supporting musical projects (CD’s)
  • maintaining VOCALOID STORE (distribution of softwares, CD’s and goodies; direct contact with the customers)

A compilation album titled “the VOCALOIDs”, featuring all the Vocaloids released so far as well as the new VOCALOID3 voicebanks will be released in September. It will have 18 tracks, and will be sold at a price of 2,300 yen (29 USD). VocaloP’s will deal with the vocal tracks, which will use VocaListener. The album comes with a “magnificent” booklet which will introduce the characters. A demo of the album was revealed during VOCALOID MIX on June 11.

The iPhone app iVOCALOID-VY1t, designed to make VY1 speak naturally, has now been officially renamed to VocaloWitter.

Studio Deen and Surfers’ Paradise will jointly organise the “i-style project” with doujin artist Carnelian doing the illustrations. It isn’t clear what this project will be yet, but Studio Deen’s managing director Noguchi talked about proposing a “new lifestyle with Vocaloid”.

MOEJAPAN has introduced the concept of a “Vocaloid you can meet”, the “fusion of a 2D Vocaloid and a 3D idol”: Vocaloidol. “Real” idols from two agencies, MOGRA and Dear Stage, will provide their voices to several Vocaloids, and they will meet their fans.

With more news coming in just a few days, it’s a great year to be a Vocaloid fan, wouldn’t you agree?

Also, I ate pizza yesterday.


Sources: ITmedia, Otanews, Goldennews, Vocaloid Otaku, Singing Robots, YAMAHA, AV Watch, Vocaloid Log.

About: kurisuto


40 thoughts on “VOCALOID3: New Features and Voicebanks (Updated)”

  1. What kind of pizza?

    Ah~ finally an article that sums up all the current events. Thank you very much.
    Now excuse me while I go fangirl with my friends.

  2. There is no limit to the amount of fangasm I am having over the CUL/Kitamura announcement. I am a huge fan of both and have loved both since their debut. September cannot come fast enough.

  3. It will have better quality audio – Yep, I swear UTAU users had it on the nail with better recording methods. VCV is very, very similar to the Vocaloid recording method, but longer samples made it more smooth.
    audio files (for off-vocal tracks for instance) can now be imported/it will now work as a VST host: YES. YES YES YES YES YES.
    It will be compatible with VOCALOID2 voicebanks: Not too shabby.
    It will support these new languages: Korean, Chinese and Spanish/The Editor will also be localized: Can I input my Korean in romaja-Revised Romanization of Korean? As much as I hate it, I'd prefer it over having to switch to the han'geul keyboard constantly. Also, THANK GOD FOR LOCALIZATION.
    Mac version “as soon as possible”: <3 <3 <3
    will actually be sold separately from the voicebanks: …noooooooooooooooooooooooooooooooooooooooooo

    [Crypton] planning on keeping their existing VB’s on VOCALOID2: Capitalizing on your current success much?
    Internet Co., Ltd will be releasing a VB for VocaloRevolution’s CUL: INTERNET YES MY FAVORITE COMPANY
    Gakupo, Gumi, Lily and Gachapoid will have a V3 update: …/dies of sheer love/ (also are they rerecording Lily please)
    The VY series will be updating to VOCALOID3: Yamaha will sell VY1 for V3 and VY2 for V3. Both of them will be sold together with the full Editor: Naturally. I love them more anyway.

  4. "It will support these new languages: Korean, Chinese and Spanish (pinyin input for Chinese and hangul input for Korean possible). "
    "The Korean company SBS Artech will release a Korean voicebank"

    YESSSSSSSSS I have been waiting for a Korean company to make a Vocaloid voicebank. I can't wait for Vocaloid3 to come out!

  5. What matters to me more than anything is that they seem to have a solid vision for where they plan to go. That fills me with alot of confidence about the future of Vocaloid

  6. I was about about to get very annoyed when I read 'It will work on Windows XP/Vista/7.' Good thing I kept reading 😀

  7. It looks like Garageband, sorta. And well, I guess a lot of modern audio programs. It looks a lot easier to use. 🙂

  8. Essentially, a lot of developement is going into the Japanese side of things but things are going to be lagging in the English side for a long time. This was not as promising as it was previously was thought for English based users of Vocaloid. We can only prey for a better pregress later on before we get left behind here entirely. :-/

    1. we are kinda of already left behind. Most Vocaloid fans hate the English Vocaloids or do not even know about them. Heck even Vocaloidism tends to ignore the news about them. Right now Prima, Tonio, Big Al, Sonika, And Sweet Ann are being prepared to be distributed by Ecapsule to Taiwan and China and barely anyone knows. Also Shu-T is helping with the advertising by making new Demo songs for the English Vocaloids new designs. So I'm not supprised that we are getting left behind it was bound to happen.

  9. The new software seems amazing- I'm also really happy that it'll be compatible with the V2 series! :DDD
    Love the future plans~ <3 <3 <3
    Oh and Crypton!! <3<3<3
    They just never disappoint me~ I was afraid that all the current V2s would've been forgotten~;;;
    Thanks for a great summary of Vocaloid events- I found this very informative and interesting-will but the album too~
    This made me really happy! ^^
    Also- Hope you enjoyed your pizza~ XDD"

  10. Ahah very good article Kurisuto, well worth the wait.

    If I understand correctly, one would have to record again voice samples to benefit from all the features of Vocaloid3, a V2 voicebank import would not reach the same quality. Therefore I'm really surprised with Crypton's choice no to adapt their voicebank to V3, especially if InternetCo intend to do it for all theirs and release a new Vocaloid like CUL who already have some notoriety. I understand that they did big investments for Miku English, but with all their licence deals they probably are the most capable of all the voicaloid companies around to support this kind of investments.
    Of course, there's also the possibility that Itoh wanted to announce some stuff during his panel at AX2011, after all CFM is probably also the only vocaloid company that can play standalone and do not need Yamaha's conferences to make an announcement. I'm wondering how they felt with Yamaha's release of a Vocaloide Label, that's direct competion to KarenT it seems.

    Anyway I'm thrilled by all this activity lately. The vocaloid phenomenom is continuing to grow and that rejoice me a lot.

    Pizza ? Quelle honte mon ami, où va la France si les habitudes gastronomiques se perdent ? ;D

    1. Thanks! 😀

      >If I understand correctly, one would have to record again voice samples to benefit from all the features of Vocaloid3, a V2 voicebank import would not reach the same quality.

      That's actually pretty vague as of now to me. Apart from the "longer samples" thing, I didn't get the impression most of these improvements they talked about will be sample-related; it seems most of them have more to do with the way the synthesis engine will process the data. For instance, hissing sounds can be dealt with easily during the recording stage or during post-production by using filters and de-essers; however, I highly doubt they hadn't already done that, so it seems it was more an artefact produced by the engine. Same for the transition between low and high voice registers: if their samples were good in the first place, they wouldn't have to re-record them. I mean, we're talking about an improvement to the Vocaloid engine, not an improvement to recording gear or techniques. So again, apart from the longer samples bit (which is very important, granted), I really think it has more to do with data processing rather than the data itself.

      >I'm wondering how they felt with Yamaha's release of a Vocaloide Label, that's direct competion to KarenT it seems.

      Oooh yeah! One of the things that struck me the most was how they're planning on shipping the full editor with the VY series, really clever move. To me it sounded like a "hey Crypton, we're here and we're planning on taking the movement back from you!"

      <s>Pizza, je te l'accorde, mais… un seul mot : "roquefort".</s>

  11. Dammit, just after I deleted CUL from my site's new dropdown menus, they have to announce that CUL is becoming an official Vocaloid. |D;;

    Wow, all this new Vocaloid stuff has me really excited. I'm kinda sad that Crypton has no plans for the Vocaloid3 engine, though. They pretty much made Vocaloid what it was today with their awesome marketing power.

  12. I find the increase in commercialization unappealing. Splitting the components of the software and charging for each is typical corporate bullshit.

  13. Mmm…..still worried about Vocalistener….does this mean it will be out in September with V3, or do we STILL not know when that's coming out? And capability with English!? YES PLEASE!! If it doesn't have it there are ways around that (as the software recognizes syllables first, words second) but it would be great to not have to use those ways. And I want to hear Engloids used with Vocalistener, we haven't heard that yet.

    Anyway, from everything I've seen, the future looks incredibly bright for Vocaloid, and this new technology seems very worth the investment indeed.

  14. This IS a good year to be a Vocaloid fangirl! (Actually I just got into it at the beginning of the year… YAY :D) OH MY GOD I CAN HARDLY WAIT! (Especially for the Gakupo and Gumi V3 updates… I LOVE YOU GAKUPO.) And an album featuring every Vocaloid? Oh yeah, I just died and went to Heaven.

  15. Actually, regarding VocaListener, how sure are you that that thing that says about their considering making it compatible with English is right? Because it seems it already is:

    You can argue about the quality, but this is definitely an English VSQ made with Vocalistener.

  16. That demo's pretty damn awesome, but I can't quite tell if the robotic tone's from autotuning or the Vocaloid itself. Still, that voice type makes me really happy and I hope she gets some use when she comes out. Also…

    "MOEJAPAN has introduced the concept of a “Vocaloid you can meet”, the “fusion of a 2D Vocaloid and a 3D idol”: Vocaloidol. “Real” idols from two agencies, MOGRA and Dear Stage, will provide their voices to several Vocaloids, and they will meet their fans."

    …this is a really, REALLY good idea and I'm surprised nobody's thought of it before these guys.

  17. OMFG! First Macne´s plugin to Garageband, then UTAU-Synth, and now… a possibility for a Vocaloid3 for Mac… WAIT A MOMENT!… If they make it… it will work on PPC macs? or only Intel ones…

    I HATE MY PROCESSOR!

  18. I have a question!~ Since the Vocaloid3 editor is being sold on Yamaha that should be a move to lower the amount of illegal vocaloid versions by selling it to anyone. Now im wondering about the voicebanks. Will the Japanese and/or Korean voicebanks be available to the overseas market?
    I don't think anyone has any answers but this is an awesome question for Japanese and new Korean companies. I would love to see them sell their voicebanks overseas.

Leave a Reply to dontmindme Cancel reply

Your email address will not be published. Required fields are marked *