Other Cryptonloid voicebank updates, collabs, & concert news (crypton_wat Twitter translations)

[Leon]

Kaito's!


I... don't hate this? It's kinda cute actually but that may be my inner Kaito bias that still stands after 10 years--

 

[Twillby]

Compared to the others he seems almost TOO calm (but I don't think that's saying much considering everyone else besides maybe Luka djlsf). I liked it more than I expected honestly but Luka is still my favorite #lukaownerbias

 

[uncreepy]

After hearing all of these, I think that the best for me was the Kagamines and the worst was Meiko. You can hear when the phonemes start/end with Miku and Kaito. I thought Luka English sounded cool, too, would like to hear a longer English clip.

Honestly, I won't give up hope completely, because 1) there's still time for them to get "fixed" (even though I feel it's unlikely), 2) factors such as a male vs female human voice as the input can drastically affect the output Cryptonloid results, and 3) the fact that you can hand tune the problem areas might mean the software can have better results than some of these clips.

Guess we go back to the waiting game after this Project Sekai excitement? Miku NT is supposed to be available mid-May (chuujun = 11th ~ 20th), Crypton isn't supposed to come back to work physically until May 20th. <--It sort of looks like it will be more like the last half of May based on these dates, doesn't it?

Not related, but anyone remember when the Miku NT page 1st got announced? They were updating the English version of the webpage in real time, so half of the page was English and half Japanese. How weird is that? Crypton sure likes to do stuff at the last minute... why didn't they just announce it after translating the page since it only took like maybe an hour or two for the English page to finish up. Did they think all us English speakers were sleeping when it got announced? Just rambling nostalgically right now, I guess...

 

[uncreepy]

To clarify, Cherry Pie isn't the name of the engine, it's the name of one of the 11ish VST vocal effectors (VST ボーカルエフェクター). Cherry Pie is specifically called a "voice analysis synthesis effector" (音声分析合成エフェクター).

And the Sonicwire page for Miku NT has been saying since December that the engine is called "Piapro Studio for NT". When it first got announced during Magical Mirai (August 30th), they temporarily called it "NEXT Generation Piapro Studio".

My translations of the initial blog announcement of Cherry Pie and NEXT Generation Piapro Studio to check back on:

Other - Cryptonloid voicebank updates, collabs, & concert news (crypton_wat Twitter translations)

The first concert for FRIDAY isn't until: 1 pm~2pm JST (10 pm CST on the 8th) and 6 pm~7pm JST (3 am CST on the 9th). (It's only 7 am in Japan right now, so no concerts have actually started, and I am not going to stay up late to find out what happens.) On SATURDAY in Japan, the concerts will...

vocaverse.network

Other - Cryptonloid voicebank updates, collabs, & concert news (crypton_wat Twitter translations)

Thank you for sharing this information with us. I didn't realize that Magical Mirai had presentation segments, I guess I just imagined people went there for the concert and maybe there was merch and displays to look at. I took a closer look at the schedule to figure out what exactly is going on...

vocaverse.network

Edit: So you can literally use new Piapro without ever touching Cherry Pie, it's optional.

 

[Trevor]

There's a good chance they hired an idol to do the original voice acting. That could easily explain the high-pitch and weird pitch bends.
It's also Meiko so there's a good chance that he will sound... wonky just to preserve her original voice quality.

 

[mobius017]

Not related, but anyone remember when the Miku NT page 1st got announced? They were updating the English version of the webpage in real time, so half of the page was English and half Japanese. How weird is that? Crypton sure likes to do stuff at the last minute... why didn't they just announce it after translating the page since it only took like maybe an hour or two for the English page to finish up. Did they think all us English speakers were sleeping when it got announced? Just rambling nostalgically right now, I guess...

Couldn't say positively about the date, but the first time I remember seeing the Miku NT page was Dec. 25, 2019. I remember following the news about preorders in down moments around Christmas stuff.

The Japanese-English updates might have been a result of one hand not knowing what the other hand was doing. Like, the Japanese side planned to make their announcement at a particular date/time, and the English side got the wrong date/time or just was held up and they weren't able to tell the Japanese side to hold off. Or else they just figured the English delay wouldn't be worth holding the whole announcement, or it wouldn't be feasible to do so if the Twitter/whatever posts were scheduled in advance or something, and so they had the English side translate like mad.

I probably missed something, but how do we know that the new Project Sekai clips use NT/Cherry Pie? Did they tell us specifically what they're using? Of the two, I guess NT would be the better candidate for just assuming it's there, but how do we know they're using Cherry?

 

[uncreepy]

They didn't announce it specifically to my knowledge, but for sure they are using Miku NT and Meiko NT, because they said they were using their new voices with them for the Pocari/Miku Symphony stuff (Meiko's talking debut for her anniversary) for them.

For Miku NT, they have been using it for every single collab talking video since 2018 (Talking: Pocari Sweat, Shinkalion, commercials for Magical Mirai stuff, collab with Sony, STV, Singing: Tokyo 150, rap collab, Pocari Sweat, etc) even though she was in ALPHA stages when it first got debuted on Shinkalion. This is definitely being used with Project Sekai and can't be refuted.

We know they are using Meiko NT for talking collabs along with Miku because Wat said so. Her Cherry Pie voice was used as early as October 2019. Why would they suddenly stop using it? I think it would be hard to say they're not using it for Project Sekai.

Luka's talking was used in a pop-up store last year in addition to singing with it at Kawasaki Jazz.

We saw in photos that the Kagamines were listed on the Cherry Pie poster presentations as usable voices to get your voice converted to. So it exists and was used privately at least, why else would it be on the poster?

So wouldn't it only make sense that ALL Cryptonloids are using NT? Each character has the same type of realistic intonation, how else would they do it? Dust off VocaListener? Tuning talkloid by hand is extremely time-consuming. I don't know why people think it's so unrealistic for them to use alpha versions of Kaito and the Kagamines and the others when they used Miku NT in her alpha stages since 2018 (it's 2021 now and even though she's in beta, they still are using it!).

The point of the NT voices is for virtual idol activities to save time tuning and so Wat wouldn't get inflammation in his hands anymore. Shinkalion Miku sounded horrible, but they used it anyway, so I think using an alpha voice wouldn't stop Crypton. They said everyone had deep learning patterns around July of last year and that they were progressing on updating all of the voicebanks from the past and each Vocaloid was mentioned at one point (ex: Wat saying he was moving on from the Kagamines and Kaito).

This is my thinking on it and nobody has to agree with me, but I think it would take a lot of convincing to make me think they aren't using the NT voices.

 

[uncreepy]

Oh, another thing to back up my tinfoil hat theory... The only person/place that announces when they use Cherry Pie is Wat's Twitter. But he hasn't updated since around February. So who is supposed to be announcing when they use it? When they put commercials on Miku's YouTube channel, it's not like they say in the video description "Hey everyone, we used Cherry Pie. Hope you like it." They just don't mention it at all. Labopton Blog only updates when they have something for sale.

So we could assume that they aren't mentioning it because 1) it's not their job because they're not Wat, or 2) they think it's already established that they are using NT. *cue X-Files music*

 

[Eviltreat16]

Oh, another thing to back up my tinfoil hat theory... The only person/place that announces when they use Cherry Pie is Wat's Twitter. But he hasn't updated since around February. So who is supposed to be announcing when they use it? When they put commercials on Miku's YouTube channel, it's not like they say in the video description "Hey everyone, we used Cherry Pie. Hope you like it." They just don't mention it at all. Labopton Blog only updates when they have something for sale.

So we could assume that they aren't mentioning it because 1) it's not their job because they're not Wat, or 2) they think it's already established that they are using NT. *cue X-Files music*

No I agree this is their NT voicebanks. If they've been using it for events prior to this, it's most likely NT. The Project Sekai samples could be older samples depending on how long they've had the voices in the Sekai game builds and where the NT voicebanks were at in development when the voice samples were made. I mean, they could've done the Vocaloids' speech earlier than around the same time as previous events from late last year but who knows? voice acting is usually one of the last things but since these are synthesizers not sure tbh

 

[mobius017]

Just going to throw this in here--I wasn't trying to suggest that it wasn't Cherry Pie/NT; I was just curious as to how we knew what it was that they were using. Didn't mean to make any kind of thing out of nothing.

 

[uncreepy]

Sorry for being over-the-top and crabby. The last few week of posts on here relating to Project Sekai made me really irrationally annoyed/angry each time something got added (too emotionally invested after having this thread since 2018 and not happy with people's reactions to them including my own) on top of irl stuff making me not want to get out of bed. It's just that it feels like whenever we add a clip to this thread, half of the people want to believe it's not Cherry Pie no matter what. On Twitter, people literally keep @ ing me in the middle of them fighting with random people I don't know about if something is using Cherry Pie or not. I feel like even though I try to explain what vocoders are, people don't understand still or don't listen very well, so I keep having to re-explain the same information over and over every time we bring up the subject on Discord/VocaVerse/Twitter. I'm getting tired of being asked about it everywhere all the time and yet people not remembering the information. Or people on Twitter who are powerful in the community spewing out misinformation about NT to their followers and I feeling compelled to correct them even though I am sick of talking about it. I wish the stupid software would come out already so I can forget about this stuff, but the more clips they post, the more worried I get like "Crud, this sounds terrible, did I waste my life following this so closely?" Sorry for ranting/being overly frustrated.

 

[Rylitah]

I think you could make a "please do not @ me if you don't have a question about something I've recently translated" tweet or something and just start linking it to people whenever they @ you to fight their arguments for them. It does sound incredibly annoying, and I'm sorry you have to go through that.

Unfortunately there's nothing you can do about popular accounts spreading misinformation -- if you haven't, you could make one final thread of all the facts you know and start linking it as well like "hey, this is wrong, please read this for actual correct info" or something kdjnfsdf. It's not much, but I think it'd work better than having to explain yourself over and over.

 

[mobius017]

I like Rylitah's suggestion about linking somewhere, either to a "'please do not @ me if you don't have a question about something I've recently translated' tweet" or to a list of all the stuff you know (I thought there was a thread/post like that awhile back, but I can't seem to find it since Dec. 25, at least in this thread.).

It's really dumb that people bug you about if something is/isn't the new software, though. Unless you were involved with making whatever it was, the best anyone can do for that question is make a very educated guess. Granted, the points you made before about intonation seem like they would let you guess pretty accurately, but still, for questions like that, they shouldn't put you on the spot for a definitive answer--they should @ the clip(s)' creator, if they really want to know. ('Course, it's likely enough that those people won't answer, but if it's that important that they know for sure, that's the chance they take.)

 

[Jikyusan]

Sorry for being over-the-top and crabby. The last few week of posts on here relating to Project Sekai made me really irrationally annoyed/angry each time something got added (too emotionally invested after having this thread since 2018 and not happy with people's reactions to them including my own) on top of irl stuff making me not want to get out of bed. It's just that it feels like whenever we add a clip to this thread, half of the people want to believe it's not Cherry Pie no matter what. On Twitter, people literally keep @ ing me in the middle of them fighting with random people I don't know about if something is using Cherry Pie or not. I feel like even though I try to explain what vocoders are, people don't understand still or don't listen very well, so I keep having to re-explain the same information over and over every time we bring up the subject on Discord/VocaVerse/Twitter. I'm getting tired of being asked about it everywhere all the time and yet people not remembering the information. Or people on Twitter who are powerful in the community spewing out misinformation about NT to their followers and I feeling compelled to correct them even though I am sick of talking about it. I wish the stupid software would come out already so I can forget about this stuff, but the more clips they post, the more worried I get like "Crud, this sounds terrible, did I waste my life following this so closely?" Sorry for ranting/being overly frustrated.

I'm sorry to hear that. I hope things get better soon.

 

[Leon]

Forgive me if this is wrong or doesn't belong here or anything, but people on Twitter are saying this is a clip with Miku NT, and if so, I gotta say she sounds good! It came out yesterday.

 

[Twillby]

My ears were so confused at first because I missed the part where it’s a duet and went “Wait this is too human and sounds nothing like Miku HOW” rofl I am not awake yet

Might be a bit of stretch to say this because we don’t have a lot of samples of NT and I’ve only heard them a handful of times but there is a certain something in her voice that makes me feel it definitely is NT

 

[mobius017]

I think Miku must be the backup singer in this video (the "na na na na na na na na na" parts). The main vocal sounds like it must be someone else.

 

[uncreepy]

What I hear:

Human (0:03-0:05)
owannai ai o daitetakunai no
motto chanto fuan ni shite yo

MIKU (0:07-0:12)
inai inai baa de enjite hoshii no
motto chanto kotaete yo

Human (0:12-0:20)
nanananananananana
"mikansei" datte nando demo iun da
nanananananananana

MIKU (0:20-0:22)
NO o karabutta ai no naka de

Human (0:23-0:26)
fugainai ai o aishitakunai no
motto chanto itaku shite yo

MIKU (0:27-0:31)
waraenai kurai ga kitto tanoshii no
motto chanto tokashite yo

Human (0:32-0:34)
nanananananananana

MIKU (0:33-37)
"mikansei" datte nando demo iun da

Human (0:38-0:41)
nanananananananana
NO o karabutta ai no naka de
nanananananananana


I think that the singing sounds more emotional than a Vocaloid can normally do, but it's still got that NT mechanical sound. The song overall sounds good, though.

 

[Leon]

^ Agreed. Thank you for the analysis! It's definitely got some good emotion to it. I wonder how NT will do on engine noise and such on release.

 

[Scarlet Illusion]

I have to say the very little bit I hear with Miku sounds nice and definitely doesn’t sound like the usual Vocaloid engine sound. Wish we could hear more of her.
(Also, anyone know who the human singer is? She sounds really good!)