Tohoku Itako’s Voice Database Development Crowdfunding Campaign
The official Greenfunding site for Tohoku Itako’s voice database development crowdfunding campaign is now available for viewing!
We were first made aware that they were planning on making a database for Itako’s singing voice synth during her birthday on June 24th. The announcement also came with little tidbits, like the Itako pocky stand donor rewards, and how they plan for Itako’s database to be similar to that of Kiritan’s, which was used for NEUTRINO.
The crowdfunding officially begins on Monday, July 6th at 7:00 PM (JST), and its goal is to reach ￥6,000,000 JPY within 28 days. Read on to see what kind of rewards you can get for supporting the campaign!
About the Crowdfunder
According to the official page, the impetus for creating this project was because Itako herself expressed that she wants to sing more! And so this project and its crowdfunding were started to help Itako get a foot into the world of singing voice synths!
True, Itako does already have an UTAU bank, but given the developments in voice synth technology these days, one can hardly fault Itako for desiring a piece of the pie!
The Greendfunding page states that they are turning Itako into an “AI Singer,” much like Kiritan’s NEUTRINO bank, known as “AI Kiritan.” NEUTRINO is a synth program that augments the special traits of the singing voice, and the below example is even given to let first-time listeners hear AI Kiritan’s voice for themselves.
By making Kiritan’s voice database open to the public in 2019, researchers like M. Morise of Meiji University took notice and AI Kiritan soon became the talk of the town among acoustics research circles.
And with the current project, they plan to follow the same pattern, releasing Itako’s voice data for researchers to use. The official page states that they currently do not know when or if Itako’s voice will be turned into an actual singing voice synth product, but they hope that someone will come along and make it happen. And with individuals like SHACHI – NEUTRINO’s creator – actually onboard as an advisor, the chances are probably good that it will happen.
Database Team
Here are the team members who will research and develop Itako’s database if the crowdfunder reaches its goal.
Database Developer: M. Morise
- Creator of Kiritan’s database
- Will be looking over all aspects of Itako’s database development, and make sure that she can sing popular songs
Project Manager: Korguchi
- Also hails from Meiji University
- Expert at phoneme research and usage
- Will make sure that anything that sounds lacking in Itako’s voice will be made better
Advisor: SHACHI
- Creator of NEUTRINO and AI Kiritan
- Will be using feedback received from AI Kiritan and use it to improve on Itako’s database
Labeling: Caparo Ulula
- Sound director for Itako’s UTAU voice bank
- Will take charge of database labeling (i.e. determining what kind of sound goes where in a song that Itako sings)
Donation Rewards
When the crowdfunding campaign goes live on July 6th, the donations options and their rewards will look like so:
￥500 JPY
- Receive an email thanking you for your support
- Receive development update emails
￥5,000 JPY
- All of the above
- Itako’s voice samples x 100
￥25,000 JPY
- All of the above
- Itako Pocky stand (can be used to play the Pocky game)
- Can badge
- Postcard
- Cloth poster (150cm x 60cm)
￥40,000 JPY
- All of the above
- Spoken recording of Itako thanking you (can address you by name)
- Clear file folder
￥60,000 JPY
All of the above
Volume 3 of Tohoku Zunko’s novel ずん好き / Zunsuki autographed by 江戸村ににこ / Edogawa Ninico, Tohoku Zunko’s official illustrator
￥100,000 JPY
- All of the above
- Cloth poster (different design from the one above) autographed by 木戸衣吹 / Ibuki Kido, Itako’s voice provider
Remember that the crowdfunding campaign will begin at exactly 7:00 PM (JST) on Monday, June 6th. Anyone interested in donating should mark their calendars!
**UPDATE ON JULY 7, 2020**
It’s official: the crowdfunder is now 100% complete and development for Itako’s new voice database will be fully funded!
But they are not stopping there. Already, new stretch goals have been announced:
Goal #2: ￥9,000,000
If reached, then Itako’s new voice database will be developed for not just singing, but also for speech! They’ll even release her lip motion and voice data for researchers to use.
Goal #3: ￥13,000,000
If reached, then similar speaking voice databases for Shikoku Metan and Zundamon will be developed as well!
Goal #4: ￥15,000,000
If reached, then the backers who donated at least ￥5,000 will get 200 voice samples for Itako, plus 200 for Metan and Zundamon!
Goal #5: ￥17,000,000
If reached, then you will receive additional 300 voice samples for Itako, Metan and Zundamon.
Concerning the above-mentioned speech databases, what will make them different from VOICEROID and other similar speech synthesis engines – which only read words and sentences that you type into the software – these new databases will be able to read lip movement and produce the proper sounds from them! Their goal is to usher in a new age of speech synthesis, and create a new way to produce output for users to enjoy!
To this end, if at least the second new goal is reached, they will develop the lip movement database along with the voice database, and present them to researchers for further development.
Imagine this: Itako reads your lip movement, and completely in real-time, she speaks and recites exactly what you said! You don’t even have to actually speak, but can even just mouth words for Itako to read and recite the words you mouthed!
This also opens up new possibilities in 2.5 entertainment like live virtual stage performances with full speech. Currently, when a voice actress recites lines for a virtual character, she must use a voice changer that will make her sound like the character. That means the voice actress must also change how she speaks to match the character’s speaking style.
But let’s say you don’t have to change the way you speak. Let’s say you just speak normally, and Itako will read your lips and not only recite what you say in her own voice, but even in her own characteristic speaking style!
These, of course, are all very lofty goals that won’t be achieved right away. But all great things have a beginning, and the organizers for this crowdfunding campaign are taking it upon themselves to make this the new beginning for the future of the voice synth world. Let’s support them in this endeavor to make the future now!
1 Response
[…] crowdfunding campaign for the development of Tohoku Itako’s singing database has succeeded in meeting its initial […]