• Hello Guest, recently a massive leak of e-mails and passwords ocurred on the internet. Please read more in here to see if you have been compromised.
  • Bookmarks are back! Read more in here.

Real-time voice conversion software aimed at VTubers called "Voidol" launches for $7.99

uncreepy

Visual kei enthusiast
Apr 9, 2018
420
USA
Source 1
Source 2
Source 3

About the software:
  • The app is going through a campaign of being $7.99 until July, even though it's true value is about $17.99. Purchase here: ‎Voidol
  • Men can easily convert their voice to female voices, women can easily convert their voices to male voices, it's high quality and has a simple interface.
  • It's developed by CRIMSON TECHNOLOGY, Inc.
  • Only the Mac version is available now, but the Windows version is planned to release on May 6th (Japan time).
  • The full name of the software is "Voidol - Powered by R.C. Voice" (Voidol - Powered by リアチェンvoice -).
  • It has been in development for a long time-- the oldest YouTube video on their channel is back from November 2017!
  • Though the app appears to be sold in English, Japanese, and Chinese, it appears the voices are only in Japanese.
  • It's aimed at VTubers so they can convert their voice in real time for things like livestreams, singing, and talking.
The voices:
The software comes with 2 free voices: CRIMMZOH (くりむ蔵) a cute monster, and Iroha Otomiya (音宮いろは) a white haired schoolgirl.
More character voice models are planned to be sold.

Introducing the voices:
Note: There might be some translation discrepancies if they ever release anything officially in English. I did my best guess at a couple character/planet names.

CRIMMZOH / CV: ? / Included for free
1946
(Probably named after CRIMSON TECHNOLOGY, Inc.) He's some sort of music alien that came to earth from Crim Planet, he likes to float horizontally and likes humming.

Zunko Tohoku / CV: Satou Satomi / Add-on voice
1947
The VOCALOID/VOICEROID we know and love.

Cutie Alien Peroro / CV: ? / Add-on voice
1949
He's an alien from Peropero Planet and is licking themed (note the suckers and the fact that his name has "pero" (lick SFX) in it.

Tsubasa Koeno / CV: Kotori Koiwai / Add-on voice
1951
(Her full name means "wings of a voice".) Yes, this is the voice of Kotori Koiwai, who recently voiced Hime and Mikoto Meika! Another familiar voice for vocal synth fans. She's a young girl with a gentle adult voice. Nobody knows much about her because she doesn't say much. She's an Aquarius.

I couldn't find their introductions, but here are the remaining voices...

Yoshida-kun / CV: FROGMAN / Add-on voice
1952
He's the main character of Eagle Talon and also a VOICEROID. Popular character in Japan and has multiple TV shows.

Iroha Otomiya / CV: ? / Included for free
1953
The pink, blue, and green are based on the input/output meter. She is not on the iPhone version, she's new for the Mac version.

Dub fairy Miranda / CV: Maya Okamoto / Add-on voice
1959
The broom microphone makes your voice into a Hollywood celebrity's. There's some confusing story about her "reel recording bag" being related to faeries and the "reel messages" getting delivered to heaven.

Tsukasa Otoshiro / CV: Takayuki Fujimoto / Add-on voice
1960
He's an ikemen (hot) boss that's longing for someone. (Note: Finally, an actual adult male ikemen voice with proper shoujo-tastic art we've all been thirsty for.) He always lights up the office with his smile and is helpful. He has a trademark yellow tie. His wonderful male voice will sooth women! (Er, okay. That's a sentence I translated with my own two hands.)

Demonstrations:

He's using CRIMMZOH.

1954



Note: It appears they are at CRIMSON TECHNOLOGY, Inc., because these demos by DTM Station are using the same bedazzled microphone. (See video below.)


They said that the Mac version had the engine rebuilt to be optimized for Mac compared to the iPhone version, so using a microphone and an audio interface creates a higher quality result compared to the iPhone version.

The app:
1956

There is a 50ms latency in the conversion of voices, so it makes you feel uncomfortable until you get used to the lag.

Yoshida-kun and Zunko have character restrictions, so before using their voices for dictation of things you want to sell, you'll need permission.

1957

1958

DUCKING lets VTubers use the weak (highlighted in blue) and strong DUCKING so it can detect when stuff like background music is playing and adjust the volume accordingly.

For INPUT, you tell the software if you are a male or female so it can adjust to converting your voice properly.

Extra Kotori Koiwai stuff:
They interviewed her and I wanted to point out the interesting stuff. It seems she found out about voidol from a DTM Station article and wants to use the software to sing with a male voice. She said she had to read about 200 sentences in a clear voice to preserve the quality, so it was a weird/different experience than normal voice acting. Kotori hasn't gotten to try using the software yet and wants to test out her character, Tsubasa.

Kotori revealed her favorite lines she had to read, which she thought seemed touching but sci-fi-ish:
「なぜなら、僕は声そのものだからです。

マイクに向かって喋るだけで、
そしたら、その子の声になってあげて、
そしたら、たくさんお友達を作って

本当は、いっぱいお喋りしたいのに、
ステージで、声を出すことが出来なくて、
その時に、使ってもらう予定なのが、
僕の声に生まれ変わって出てくるらしい。
顔も体も、まだ見たことがありません。


でも、僕の声を聞いてもらうことが出来るので、
これから新しく出会うお友達に届けられるって、
世界中の、たくさんのお友達に」
It's basically saying stuff like "I will face the mic and become this character's voice, I want to make many friends throughout the world, I have no face or body, I am the voice itself" that type of stuff.

Final thoughts:
Here is a demo on the YouTube channel of CRIMSON TECHNOLOGY, Inc. to hear the voices in isolation (so not being demonstrated in real time, just the result). I think it is cool but sometimes sounds wonked up (but it's hard to tell cause many videos are from a year ago and might not reflect the quality of the Mac version). I had no clue this was being developed for so long and want to try the Windows version. I can't believe this unknown company released a voice conversion software sooner than Crypton's CHERRY PIE or CeVIO or any other deep learning based software.
 
Last edited:

GreenFantasy64

カイミクfangirl
Apr 9, 2018
364
soundcloud.com
Thanks for translating it. :mikiv4_ani_lili:

He's an ikemen (hot) boss that's longing for someone. (Note: Finally, an actual adult male ikemen voice with proper shoujo-tastic art we've all been thirsty for.) He always lights up the office with his smile and is helpful. He has a trademark yellow tie. His wonderful male voice will sooth women! (Er, okay. That's a sentence I translated with my own two hands.)
:clara_ani_lili:
Also, in the videos, his voice sounds a bit robotic.....actually reminds me of Kaito V1.... :mirai_ani_lili:
 
  • Like
Reactions: uncreepy

uncreepy

Visual kei enthusiast
Apr 9, 2018
420
USA
Here's a video of a guy demonstrating Iroha's voice without us having to hear his voice at the same time. I think it properly shows the actual quality of the software.


Obviously, there's some crackling and weird parts, so it's not perfect. I'm curious if they are going to keep updating the quality or or if this is it. But I feel like it's so cheap that when the Windows version comes out, I really want to try it, even if it's not perfect. I want to feel what it's like to have Tsukasa's voice come out of my mouth and try singing. LOL

I'm wondering if when Crypton releases their version if it will be "better" because you can edit the tuning to get rid of errors like this through the Vocaloid software, while Voidol is just uneditable (for the most part) voice conversion.
 
Last edited:
  • Like
Reactions: GreenFantasy64

Mika

crying over Rune Factory 5
Jul 5, 2018
117
21
Germany
That's such a cool idea! And I can totally see people buying it to make song covers with Zunko without having to tune her haha. I mean, if you are a relatively good singer it will be way easier and cheaper than Vocaloid.
I think I will buy it too when the windows version comes out, seems like it can be a lot of fun to use.
 

Wario94

Aspiring Fan
Jan 5, 2019
29
24
Sounds interesting!
Just curious, besides Tohoku Zunko, is it possible if very familiar Vocaloids like the Crypton family and Meika Hime and Mikoto would ever appear on both Voidol and R.C. Voice?
 

uncreepy

Visual kei enthusiast
Apr 9, 2018
420
USA
Sounds interesting!
Just curious, besides Tohoku Zunko, is it possible if very familiar Vocaloids like the Crypton family and Meika Hime and Mikoto would ever appear on both Voidol and R.C. Voice?
Meika Hime and Mikoto would not appear on Voidol/R.C. Voice, because their voice provider (Kotori Koiwai) is already on the app for the Tsubasa Koeno character. I doubt the Cryptonloids would be on here, either, because Crypton seems to be unwilling to have their characters on other software (like text to speech software, and they went as far as making Piapro over Vocaloid Editor, so... yeah). Plus, they have been working on their own voice converter (Cherry Pie), which is probably going to be superior in quality to this app, so I highly doubt they would bother reaching out to each other.

Tohoku Zunko (who has a Voiceroid and Vocaloid version) and Yoshida-kun (who has a Voiceroid version) are special, because Zunko is used to promote tourism, I think Yoshida is supposed to promote education or something, and both Zunko and Yoshida have special licensing agreements related to using their voice/design for profit.
 

uncreepy

Visual kei enthusiast
Apr 9, 2018
420
USA
Chinese language Voidol (interface and voice(s)):
Voidol is planning on expanding to the Chinese market. The vocal coating technology will be able to convert Chinese phonemes in the Windows version of Voidol (which was supposed to come out in May, but it's late). They are currently looking for a partner to help them get a location to exhibit the software in October. They said they have the voice of Chinese-language voice actor(s), but I'm not sure how many voices they are offering or what the character(s) looks like.

Source


Livestream demonstrating the power of Voidol in about 24 hours:

Also, DTM Station is doing a livestream with Crimson Technology to promote Voidol and will be doing talking and singing demonstrations in about 24 hours.
One of the guest singers will be Freddy Hatae from GUEEN, they are the vocalist of a Japanese QUEEN tribute band. I know that sounds weird, but they demoed a Voidol "Bohemian Rhapsody" in April:

When it says "Voidol: OFF 声 (raw voice)", the man standing up by the silver mic (Freddy Hatae) is singing, when it says "Voidol: ON" it is coating his voice with the "Bohemian" voice model. (I don't know if this can be used by the public, but I suppose I will try to watch the livestream to find out, also because I'm probably the only person on the English side emotionally invested in Voidol and have been waking up to disappointment every day Voidol still isn't out). Also, doesn't this mean that Voidol is clearly capable of English?!

The second singing guest is Yuki Shinohara from GLIRUS.

Correction:
I noticed I read the date on one of the initial sources I cited incorrectly, I thought they had a promotional sale for some of the characters (Yoshida-kun, Miranda, and Tsukasa) to be half off, but the article was actually a year old and was for when R.C. Voice first got released. I got confused, because the article was posted at around the same time the PC version came out. So, to my current knowledge, there isn't a sale for those three characters (though the Voidol app IS on sale at a discount). I'm sorry for the confusion and removed the mentions of the expired promotional price for those three characters.
 
Last edited:
  • Like
Reactions: mobius017

uncreepy

Visual kei enthusiast
Apr 9, 2018
420
USA
pre-post tl;dr > Voidol Windows version coming by end of June for $20. Confirmed angel girl's voice = Koeno Tsubasa and hot office dude = Otoshiro Tsukasa. I took a screen shot of unreleased character voice names (scroll down till you see :alien:s).


Okay, I watched the livestream and am back to provide my humble report on it.



I thought DTM Station was gonna exclusively talk about Voidol, but no... it was a 2 hour livestream and they didn't talk about it until 1 hour in. (Will summarize all topics mentioned for those interested. Plus, I sat through all that and don't want to have wasted an hour of my life. And extra plus, this is my thread and I can do what I want.)

The two hosts are Ken Fujimoto (the wonderful guy who writes all the juicy vocal synth news that I translate into English) and Aki Fumitada (not sure what to say about him).

1) 0:00 - 23:00 It started with Aki and Ken showing Kyouka Tamura (I guess she works for DTM Station but is also anime industry-related) how to use Studio One for a whopping THIRTY MINUTES. They would show her how to press play and how to speed up the tempo, so then when it was her turn, she would go "Ah! It's moving!" and be like mega surprised that it was moving faster. They also showed her how to move notes around and other non-Voidol related stuff. (I have a hard time believing she's never used Studio One before, if it were me, I would have "new software rage" after 5 minutes.)

~cue several minute long "DTM Station Plus" song~

2) 28:00 - 38:00 Next, they talked about the Satou Sasara and IA deep learning song. Admittedly, I was translating and not paying attention, but Aki kept asking why IA was singing in English and they even had the sheet music with the "deep learning~" lyrics printed out.

3) 38:00 - 40:00 Talked about a concert a "Live Stage Guilty" in Tokyo. (Didn't pay attention.)

4) 40:00 - 50:00 Talked about Native Instruments and KOMPLETE AUDIO sale stuff and a small recording device and maybe some other stuff (might have fell asleep during this part).

5) 50:00 - 55:00 Demonstrated the new changes to Haruno Sora's attack and release effects. You can hear the stuff they showed off on their article cause it's quicker to load and more to the point: VOCALOID5の桜乃そらがアップデート。これまでボカロでは不可能だった歌い方が可能になり、結月ゆかりや初音ミクにも適用可能!|藤本健の “DTMステーション”

6) 55:00 - 58:00 Talked about MIDI 2.0's new specs.

7) 58:00 - 2:08:00 And drum roll... Finally the Voidol stuff!

After self introductions, Freddy Hatae explains all about Voidol and says that R.C. voice > real chain voice > got its name from real(time) voice change(r) (this is the simple iPhone app version of Voidol). He explains how the tech coats your voice and explains about deep learning and all that jazz. They are using the Mac version to demonstrate for the stream. He reveals the Windows version is coming out by the end of June (yay) and said it is expected to cost 2000 yen ($20) (they might do a sale for a limited time, though).

They take turns demonstrating the voices. Apparently, Miranda is supposed to be like Meg Ryan. I confirmed that the angel girl's name is "Koeno Tsubasa" and the ikemen office guy's name is "Otoshiro Tsukasa". The guys kept laughing at Yuki because her real voice sounds like CRIMMZOH's. It seemed like the voice conversion worked better on Yuki's female voice (less wonky random sounds) and that female speaker + female voice or male speaker + male voice works better.

When they would switch the character voice, I noticed that you could see a giant list of the unreleased voices:
2028

I THINK the list says the following (I put an alien emoji by the super secret voices, they didn't demonstrate these and I don't know what language they will speak or if they will be released):
:alien: Bohemian voice =showed on previous post, sounds like Freddy Mercury
くりむ蔵 = CRIMMZOH
音宮いろは = Iroha Otomiya
吹替の妖精ミランダ = Dub fairy Miranda (CV: 岡本麻弥 / Okamoto Maya)
キューティ・エイリアン ペロロ = Cutie Alien Peroro
声乃ツバサ = Koeno Tsubasa (CV: 小岩井ことり / Koiwai Kotori)
音城ツカサ = Otoshiro Tsukasa (CV: 藤本隆行 / Fujimoto Takayuki)
鷹の爪団 吉田くん = Eagle Talon Yoshida-kun (CV: FROGMAN)
東北ずん子 = Tohoku Zunko (CV: 佐藤聡美 / Satou Satomi)
:alien: ペッパー = Pepper (male or female, we do not know)
:alien: pep001x_target_1.88 (sounds like it must be a code name for "Pepper"?)
:alien: 女王シャッフル = Queen Shuffle (er... what?)
:alien: ジャックブロウ or プロウ = I can't tell it it says JAKKU BUROU or JAKKU PUROU (So I guess it would be "Jack Blow" or "Jack Plow"?)
:alien: lfls002x_target_0.7 < I can't tell if those are L's or i's or something else. No clue

They finished up the stream by showing the wonked up motion capture for Koeno Tsubasa that uses Adobe Character Animator if you want to be a VTuber of her (the art of her is free, but you have to have the software): リアチェンvoiceで人気の声乃ツバサ (CV:小岩井ことり)でV-Tuberはどうですか? (Why not Live2D?)
2029

Bonus: One of the people who commented on the stream commented with "Don't progress babiniku" (when guys have female avatars for VTuber vids but have a male voice... until the Windows version comes out, that is), made me laugh out loud. Also, the iPhone version can only be used with 7+ for R.C. Voice.
 
  • Like
Reactions: mobius017

mobius017

Aspiring ∞ Creator
Apr 8, 2018
534
The two hosts are Ken Fujimoto (the wonderful guy who writes all the juicy vocal synth news that I translate into English) and Aki Fumitada (not sure what to say about him).
And Aki's like: orz
:)
1) 0:00 - 23:00 It started with Aki and Ken showing Kyouka Tamura (I guess she works for DTM Station but is also anime industry-related) how to use Studio One for a whopping THIRTY MINUTES. They would show her how to press play and how to speed up the tempo, so then when it was her turn, she would go "Ah! It's moving!" and be like mega surprised that it was moving faster. They also showed her how to move notes around and other non-Voidol related stuff. (I have a hard time believing she's never used Studio One before, if it were me, I would have "new software rage" after 5 minutes.)
Yeah, that sounds acted-up. I mean, they said it'd move faster, right?

I'll probably never be a vtuber, but the release date is good news for those who want to do that!
 

uncreepy

Visual kei enthusiast
Apr 9, 2018
420
USA
I'll probably never be a vtuber, but the release date is good news for those who want to do that!
Personally, I don't think I'll become a VTuber any time soon, but I really want to try out the vocal coating technology. It's similar to Cherry Pie (but probably a fraction of the price), so it seems like it'll be good practice for when that eventually comes out. I have some ideas for tests I want to do when I can purchase Voidol and will probably share them on here. (I know not a lot of people are interested because they have no use for it, like, they can't sing or they can't speak because they don't know Japanese.)

Kind of weird (okay, very weird), but I have been a huge fan rock/metal for tons of years and have been translating so many Japanese songs and sang along with them in order to learn Japanese, when I would sing I just feel horrible like "Why can't I have the voice of a cool guy come out of my mouth instead of my girl voice?" I just felt so strongly that I love men's singing voices that no matter how good I sang, it wasn't "right". So seeing that Voidol can be used for singing and talking and they have male voices available (especially Tsukasa) is making my mind explode waiting to get my dream achieved of having a male voice come out. I never knew such technology could ever exist, even if it's not perfect. Hopefully the recordings can be edited with something like Melodyne in order to make the singing/talking sound more perfect, maybe it could be a fun hobby to sing covers with the Voidol voices. (I guess I'm gonna be the equivalent of those VTubers with cutesy female avatars and male voices, but the other way around.)
 

uncreepy

Visual kei enthusiast
Apr 9, 2018
420
USA
I wrote a chart with ideas for what I want to test when Voidol comes out at the end of the month. I want to test it out in various situations, such as speeds, talking styles, having background noise, different types of pitch, pre-recorded audio being played out loud, etc. If you have anything you'd like to be tested, please let me know.

Here's a picture of a portion of the chart:

Edit: Forgot to mention, but I was checking out the Voidol website and it sounds like it will only have Iroha and CRIMMZOH for a while, the other voices will come later? If this is true, I am really disappointed! I wanted to try out Tsubasa and Tsukasa! Why would they show off all the voices if we can't use them yet? How cruel!
 
  • Like
Reactions: Mika and RazzyRu

Users Who Are Viewing This Thread (Users: 0, Guests: 0)