Crypton Future Media Presentation on Their New Voice Synth Engine
At about 2:30 PM on August 31, 2019 during Magical Mirai 2019 – and during Miku’s 12th anniversary – Wataru Sasaki and Hiroyuki Itoh of Crypton Future Media, and Masataka Goto of the National Institute of Advanced Industrial Science and Technology (hereafter “AIST”) held a panel to discuss their future plans, and the future of the Cryptonloids in general.
Alongside announcements like their ongoing collaboration with K!!te and Songle Sync, and a free update to Hatune Miku V4 Chinese sometime this year, the two companies also announce the new engine and software that they are developing for Miku and the rest of their character voice banks. The below information is gleaned from tweets posted by NAT and Cyross on Twitter, who were present at and shared photos and information from the presentation.
New Voice Databases
In one of the first slides of the presentation, Sasaki describes Crypton’s aim, which is to create new voice databases. What they want is for each of their virtual singers to be able to sound great no matter which pitch they sing in, and no matter what syllable they are pronouncing. This means being able to maintain sound balance during syllable transition, keeping the consonant part of each note clear and concise, and resonating each vowel sound.
To meet this aim, Crypton has begun the process of adjusting Miku’s entire voice database. Once Miku’s new database is complete, it will serve as a reference for creating the new and improved databases for the rest of their virtual singers.
New Voice Technology in Development
Sasaki then moved on to the new voice effectors that Crypton has been developing for about two years. They plan to implement them in piapro studio for use with each of their virtual singers’ voice. These new effectors include VocalDrive and CherryPie, which were introduced earlier this year as a project by Crypton’s technology research lab, Labopton.
In fact, we have recently heard one application of VocalDrive with Mitchie M‘s Seraphim On The Ring, which uses the effector for KAITO’s spoken parts. Through this, we can get an idea of how VocalDrive and the other effectors in development can make a character voice bank sound.
New Voice Software Series
In the next slide, Crypton reveals that it’s not just new voice effectors, but an entire new sound system that they will be using for their virtual singers. Perhaps the biggest surprise is that it is not with VOCALOID5, but with an entire new software series that they are developing with AIST. Both Itoh and Goto join the panel at this point.
The flow of the new sound system would be to take the base phonemes from the CVVC databases, run them through piapro studio editor where they can be read, then have them further processed and tuned in the voice effectors before being put out.
The development goal of the new sound system is to preserve what makes each character voice sound unique while giving the users more control over their individual nuances. Their aim is to develop a next generation E.V.E.C. and new voice expression technology with various voice manipulation capabilities.
Continuing talks on their technology in development, the panelists went on to describe the next generation piapro studio. The slide mentions that it will come with many improved functions and features such as:
– Wave display
– Pitch display
– Semi-automatic parameter control
– Pronunciation nuance control
– Detailed character sound customization
– Multiband formant shifter
– Voice tension (“harmonic pattern”) adjuster
– Vocal drive
Crypton and AIST plan for a release of the currently nameless voice software series sometime during the first half of 2020. Even after the final product is complete and released, they plan to continue developing and updating it as time goes on.
Remarks from the Panelists
According to NAT and Cyross, Itoh explained that the decision to develop the new sound system, engine, and software series for the future of their character voice banks is one that was made after talks with YAMAHA, creators of the VOCALOID engine. Their top priority is to support the existing users of their products, and will market the new software in a way that will make the transition easy for customers. Itoh adds that they will be working with individuals like Akira Nakamura of AIST to meet their company goals, but also wishes to continue working with YAMAHA in some way.
This means that the piapro characters will not be VOCALOID products going forward. Itoh explains that Crypton Future Media is a sound company whose singular goal is to use the power of technology to bring new sound experiences to the future of virtual music and its fans. Using VOCALOID up to now for their character voices was an extension of that goal, and development of Hatsune Miku no doubt played a big part in the growth of the VOCALOID community and culture, a culture that Itoh respects and loves greatly.
In the end, though, they are still a sound company that is ever changing and looking for more ways to support creators and producers; their collaboration with K!!te and development of their own engine are examples of this.
Sasaki then takes the mic to sum up the main points of the presentation, and assures listeners that Miku will still sound like Miku even on a different engine. He also explains that the main reason for not going with VOCALOID5 for Miku was because of problems they discovered when running Saki Fujita‘s voice in the engine; it was not because of some kind of dispute with YAMAHA. Crypton will in fact proceed in the most appropriate fashion possible so that their departure from using the VOCALOID engine does not jeopardize their relationship with YAMAHA, and will continue to work with them.
Based on Cyross’s observation, Sasaki may have even shed a tear at one point during the presentation.
In conclusion, Itoh remarks that he very much likes the vibe of the new deal that they have with AIST and the technology that they are developing, and looks forward to a long working relationship with them.
Other VOCALOID-Using Companies
As for how Crypton Future Media’s departure from VOCALOID will influence other companies who still use VOCALOID for their voice banks, we will have to wait and see how things play out. There has been one interesting tweet from Tohoku Zunko’s official Twitter account that suggests Crypton may not be alone in researching and developing new engines for future releases of their character voice banks.
In the tweet, she brings up how the Tohoku Sisters have software that are compatible with not only VOCALOID, but also UTAU, VOICEROID, and even voidol, and they may not be stopping there. They are currently preparing the reveal of heir own new singing voice synth engine data for Tohoku Kiritan with the help of Masanori Morise, an engineering doctorate at Meiji University.
It is probably safe to say that the tweet was prompted by the news from Crypton Future Media. An important distinction to note is how they, too, seem to be developing their own patented singing voice synth engine as opposed to changing over to an existing engine (such at 1st Place to CeVIO Creative Studio, or Quadimension to Synthesizer V).
What the future holds for the VOCALOID community – or even if it will continue to be called that – we will keep a close eye on and will always continue to report.
We thank NAT and Cyross for their tweets and information sharing!
Thanks for the article! I’m a little worried how this change will impact producers and their work, but I’m hopping this will end well.