• This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn more.

1 year's worth of crypton_wat Twitter translations

OP
OP
uncreepy

uncreepy

twins: Japanese translator Poobah & 3D Poobah
Apr 9, 2018
86
120
USA
#21
Len Kagamine's been added to the lineup this Friday. Last week appears to have been skipped due to Golden Week.

This entire week was a bloodbath! But, tomorrow I have to wake up and watch Shinkalion in real time!! After that, it’s dinner with the people at Bilibili!!!
Notes:
He’s probably referring to the chaos of Golden Week.
Shinkalion airs today at about 5 pm US time.
I wonder what part of the April 30th memo Bilibili is related to. Could it be about Miku's Chinese voicebank?

Notable reply:

kazu2:
When the story about Black Shinkalion is over,
will Miku-san have a break for a while?
(Shinkalion is interesting, so even if Miku-san doesn’t show up, I think I will watch it, though)

crypton_wat:
Things like how long the space is open, I’m not sure if I’m in a position where talking is bad, as a matter of fact, I’m itching to talk! By the way, even next week it’s planned to adjust Hatsune Miku’s (from Shinkalion) sound source.
--
Note: Wat's April 30th memo said Miku would be in Shinkalion again and that it was coming up fast.
 
OP
OP
uncreepy

uncreepy

twins: Japanese translator Poobah & 3D Poobah
Apr 9, 2018
86
120
USA
#22
Several new tweets today. The first (see "ORIGINAL TWEET") was deleted and "REPLACEMENT TWEET" replaced it after about 20 minutes. Too bad for Wat that I get text alerts and had already saved screenshots. I've highlighted the sentences in order for us to easily compare the changes to his wording.

wat_deleted.png
Note:
* (Domo) yoroshiku onegaishimasu: Kind of untranslatable, but used when asking favors and showing thankfulness. Please be kind to me, thank you for understanding, please do me a favor…

Both say he wants to increase the human-ness of Miku, but he either 1) accidentally revealed when Miku will appear on Shinkalion after the Black Shinkalion arc (2 months from now), or 2) he wanted to make it seem like he didn't have much time to accomplish this task. Reading comments he's received, some people praised him for his hard work and how cute Miku sounded, or were disappointed that she sounded like a machine (that's what "from all sides" means, critique and praise).


His note about what happened to the original tweet:
The tweet from some time ago, even though I received RY & replies, it disappeared with the retweet, I’m very sorry m(_ _)m
The way he wrote this makes it sound like it just vanished. It wasn't 消えた (vanished), it was 消した (deleted)!


Clarifies that less engine noise is compared to CRYPTON products:
Maybe V5 isn't coming? Or at least what they are doing is not related to Yamaha's methods (more like UTAU's recording methods of short audio)
Thanks for so many opinions! I’m sorry I can’t reply individually, but the current state hereafter: I think we can decrease machine-ness to a degree of 1.5-times※, (※compared to our company, Crypton's, products).

Also, the voice for this month has already been delivered, and is not included in this promise. I am very sorry, but from now on, yoroshiku onegaishimasu.*
Notes:
Based on the comment, it seems “the voice for this month has already been delivered” is talking about Miku’s speech from the Black Shinkalion arc of Shinkalion.
I have no clue how they can measure the human-ness vs. machine-ness to get a number like 1.5x. :S


The machine-ness he's talking about:
Currently moving from Chou Kabuki's recording location. This year isn't to the degree of Shinkalion, but with the current VOCALOID engine, the number one thing the voices are poor at is how there is a minute (fine) roughness to the overtone/harmonics (aka engine noise), it seems like the voices will become a bit more pure/undiluted than in the past. *It's not that I eventually want to make Miku human-ish, I'm thinking of it as a demonstration test/experiment concept that we should accomplish.

No "currently pinned tweet" anymore:
Wat's "currently pinned tweet" is gone. He has no pinned tweet anymore. This is what it used to be:
Currently developing a semi-automatic tuning* program, it will be a little while longer before we can do a demonstration test. We are anticipating a release that's accepted by beginners as having easy-to-use features, but it seems it will have various practical uses! *excited* (´゚∀゚ノ)ノ
Note:
Tuning is when you use parameters like BRE and DYN to make the vocals sound good.


What do you guys think all these mysterious actions mean?
 

WintermintP

Lead Guitar, One Minute Winter
#23
Actually, tuning is when you not only automate parameters but also create note transitions so that the slurs can be a lot smoother/sharper depending on the effect you want. There's a lot more to tuning than just parameters.

All that aside, it seems to me that he's just trying to keep everyone's hopes up when it comes to the Append hype, but my two pence are that I'm definitely against the idea of Appends. I mean, they give you more variety in a bank, and that's great, but the problem is that Appends are probably extra charge, and going to be a problem for people who already bought the base 5 banks like I did (unless they're going to provide those Appends for free and I don't know about it). Also, I'm concerned regarding the V2 voices, with the main problem being the physical quality of the recordings. Because we didn't have all this advanced technology back then, I have no clue how good the V2 voices will sound in a new engine. I should also mention, we have new features such as Growl that didn't exist back then, so... what's going to happen with that?

Also, my hunch is that a better Miku voice might not be happening. I don't know about Wat specifically, but I've seen the pattern with Japanese companies long enough that I can't help but be skeptical every time they seem to try to keep everyone's hopes up.

Anyway, it'll be a little sad to see that Wat doesn't want to make the voicebanks sound more human, but I have no clue if any developer actually wants to make more human-like voicebanks. If one comes up and that becomes a reality, I'd be the first in line to buy that bank.

WintermintP
 
Last edited:
OP
OP
uncreepy

uncreepy

twins: Japanese translator Poobah & 3D Poobah
Apr 9, 2018
86
120
USA
#24
Yesterday, when having a preliminary meeting with the people at STV*, I was asked "My kid likes Negimiku! Hey, now I understand Negimiku, don’t I?!!", with the flow of 10 years, even "Hachune"'s name got quite deformed.**
Notes:
* Based on quick research, “STV” meaning “Shanghai Television (上海电视台)” seems to make the most sense, due to Miku having a Chinese voicebank.
Wikipedia: Shanghai Television - Wikipedia
This is their official YouTube channel: SMG上海电视台官方频道 SMG Shanghai TV Official Channel

**Hachune = Hachune Miku. Hachune's name got deformed = somehow ended up becoming Negimiku. Negi = leek.

This tweet leave more questions than answers. I guess he's talking to Shanghai TV cause the most logical thing would be for her to have a televised concert, possibly with her Chinese voicebank or something. His next tweet (which I'm not bothering to translate) is him talking about Miku's leek and Hachune Miku, but I feel like this particular tweet was a "subtle" namedrop of STV. Having someone call Hachune Miku "Negimiku" implies that the person they spoke to understood Japanese, rather than speaking to a translator (seems too obscure for a Chinese-only speaker to know a random Vocaloid's "deformed over time" joke-character name (Hachune vs Negimiku), doesn't it?

Actually, tuning is...
Yeah, I just put a quick note there to simply define it, cause not all Vocaloid fans own a Vocaloid or know how the software works. I'm a wanna-be translator, not a dictionary!
...the problem is that Appends are probably extra charge, and going to be a problem for people who already bought the base 5 banks like I did (unless they're going to provide those Appends for free and I don't know about it). Also, I'm concerned regarding the V2 voices, with the main problem being the physical quality of the recordings. Because we didn't have all this advanced technology back then, I have no clue how good the V2 voices will sound in a new engine. I should also mention, we have new features such as Growl that didn't exist back then, so... what's going to happen with that?
Wat said he wanted the Appends to be affordable and said some features would be free in a really old tweet (so things could change). I don't know how it would work, would you have to own V4 in order to get the Append? Or would it also work with V3? Very confusing the more you think about it and also mysterious.
 

WintermintP

Lead Guitar, One Minute Winter
#25
Wat said he wanted the Appends to be affordable and said some features would be free in a really old tweet (so things could change). I don't know how it would work, would you have to own V4 in order to get the Append? Or would it also work with V3? Very confusing the more you think about it and also mysterious.
Based on how things normally work when it comes to generations, I'd imagine, yes, you would absolutely have to buy the V4X because support for V3 must've pretty much stopped at this point. Unless there's something going on that I don't know about.

WintermintP
 
OP
OP
uncreepy

uncreepy

twins: Japanese translator Poobah & 3D Poobah
Apr 9, 2018
86
120
USA
#26
Based on how things normally work when it comes to generations, I'd imagine, yes, you would absolutely have to buy the V4X because support for V3 must've pretty much stopped at this point. Unless there's something going on that I don't know about.

WintermintP
Sonicwire still sells V3 (Miku V3 for example) and it says:
Technical Support
We accept inquiries about the product in Japanese only.
With regard to English Installation Guide, you can check it on the page of Product Support Information.
We recommend you to purchase the product from distributors listed below if you need Technical Support in English or other languages.
So it seems like it's still relevant. I think only V2 and definitely V1 are "use at your own risk" level of zero technical support.
 
May 2, 2018
17
21
#27
I don't think the append would be for v3 Miku at least. Watt's main goal is to get you to update Miku as often as possible. It wouldn't make sense to produce low-price or free features for an outdated software.
 
Likes: WintermintP
May 2, 2018
17
21
#29
I second this. I really can't help but suspect that the new Appends won't be cheap either.

WintermintP
Crypton is odd. They released a free update for Rin and Len once, and the v2 appends were cheap as dirt considering it was a time period where each vocaloid was 200 dollars with one voicebank. (I seriously hope we get a new, free Luka bank out of this. Her moe thing is just a no in my book)
 

Users Who Are Viewing This Thread (Users: 1, Guests: 0)