CeVIO CeVIO Discussion Thread

[KingPapillon]

AHS only owns Yuki, Kiyo, and maybe Maki of that group fully, and Yuki on another engine is impossible so it's really just Kiyo, Maki, and Kou.

Wonder who it was that greenlit the Vocaloid 4 updates of all the classic V2 AHSloids then. Obviously the developers and distributers of the voicebanks themselves have some sort of pull in all of this, even if they don't outright own the rights to the character itself. Otherwise, we wouldn't have gotten Miki, Iroha, Yuki, Kiyo, and Yukari V4 all within a really short span of time, and with new voicebanks recorded for some on top of that. (Unless something has drastically changed since then)

 

[Rylitah]

AHS only owns Yuki, Kiyo, and maybe Maki of that group fully, and Yuki on another engine is impossible so it's really just Kiyo, Maki, and Kou.

Sora as well.

As long as you to sing and talk on the same render window, saves so much effort than opening 2 windows on different programs for singing and talking. They showcased this with Kiritan on the AHS video and Sasara. Is that possible in voiceroid and other softwares? or I'm just that ignorant in this kind of stuff.

CeVIO is the only one that can do singing and talking at once still (well, the only one that specializes in doing both rather than doing workarounds for other synths). This might change with A.I.VOICE next year since we know they want to do singing voices as well, but we don't have any details on that or the knowledge of if it'll be on the same program as the talking voices at all.

 

[KingPapillon]

Sora as well.

I forgot she existed...
thats how little I think about V5 lool

 

[YOYo_MAMA]

CeVIO is the only one that can do singing and talking at once stil

I can already see the multiple Let's play videos in Nico nico where Yukari will burst into singing in the middle of a long commentary talk on Minecraft or something.

 

[___]

As it was said that's how you'd imagine Yukari on good old Cevio...so...a downgrade as harsh as that sounds...

I'd argue she sounds worse than Kafu type A

 

[ruffe]

Well... I'm sold! I actually adore her, despite the usual CeVIO engine noise. Her silhouette gives me the impression that her design will be pretty elegant and I love that. Can't wait for Flower.

 

[Stardust]

I gave up on Kafu, but now I feel like I should give up on AI as a whole and hope that Pro will be better. Or cave and buy the CeVIO we have now.
Perhaps it is just a poor demo, but, it sounds like Yukari V4 with the CeVIO noise. Nothing about it sounds different or smoother to me like I thought it would. As bad as Kafu Type A sounds, I could at least tell that it was an AI bank. I do not know how to explain it, but something about how she sounds is different enough that I could tell it was not from a sample-based engine.

The only plus I can see from this is a new design for Yukari, and that it may be cheaper then Yukari V4, so if you want Yukari you could buy this version instead.

 

[Blue Of Mind]

I think I need to hear more demos before I make a firm judgement. Right now, I think Yukari at least still sounds like her Vocaloid self, just with CeVIO's infamous engine noise. She's a massive step up from Kafu at least.

Her silhouette looks interesting though. It's making me wonder if they've taken an approach similar to Gumi English's design by making her wear a prom dress.

 

[YOYo_MAMA]

Yukari's weird sound may just be a tuning preference by the producer. A Cevio AI demo with Sasara was uploaded and Sasara sounded way better than her previous voice. This is basically the same song from the demo they released a year ago but we can see the interface at work.

 

[YOYo_MAMA]

Here's another video but Sasara swaps between a realistic sounding and her noisy "old" voice.
I wonder if all Cevio AI VB's can have this feature, then there's hope for KAFU type A to exist after all.

 

[lIlI]

Yukari sounds nice so far! Not as smooth as her Vocaloid version, but much less noisy than previous CeVIOs. (Although it's hard to make much of a judgement on one short clip.) It's a shame the noise is still present (very noticeable in Sasara's demo), but it will be cool to have another way to use Yukari.

 

[YOYo_MAMA]

After the disappointment that is KAFU and Yukari (not really a disappointment but some people do) . This Hype is really setting people's expectations way to high to be reached. With a competitor like SynthV AI, any changes when it's not even released yet just turns people off.

 

[lIlI]

Tinfoil hat on: Yukari sounds like she was trained with Vocaloid samples, not her actual VP. Perhaps the company wanted her to sound as close to the original product as possible, or they didn't want to spend extra money recording a session with a real singer.

 

[xuu]

I actually really like Yukari's voice... she sounds like her Vocaloid version, which was already very good, but also slightly more human? It does seem like they used her Vocaloid samples, or at the very least got Chihiro to record in a similar fashion. She sounds very soothing, definitely better than KAFU, that's for sure.

I think at the moment it's Sasara > Yukari > KAFU, but maybe Yukari will sound better when we get a raw sample. Looking forward to a flower demo now.

 

[uncreepy]

I feel weird about her design, it kind of looks like she has a ponytail and a prom dress. The kanji for rei means beauty, so maybe they are glitzing up her design for this. It's weird because her design for all of her versions has remained the same, why so special for CeVIO AI?

I think the demo sounds fine, but it's not super awesome/making me feel like I have to buy CeVIO AI. I hope we can hear more demos soon. Also, wasn't expecting a delay for CeVIO AI, but oh well. Looks like a lot of us aren't keen on buying it so far anyway.

 

[peaches2217]

Honestly, the silhouette is the only thing I really like; I’m excited to see how her design has changed! But as it was mentioned up-thread, she’s just... Yukari, but on a different engine. Redundant, with nothing new to offer except the buzz of the CeVIO engine. Maybe that sounds pessimistic of me, but I was hoping they’d try taking advantage of the engine jump and optimize her, and I’m sad that they didn’t.

 

[___]

I do think it's very obvious her vocaloid samples were used, no way she'd sound like that if her vp provided brand new samples for her Cevio bank, the "ai tuning" is way more minimal than Kafu's or Sasara's, she very much sounds like untuned, raw, vocaloid Yukari or at least very minimally tuned plus the buzz.

And...companies sure are hopping on this AI trend but refusing to commit to what it entails.

 

[lIlI]

Assuming they used Vocaloid Renders to train the AI (Yukari was a realistic voicebank already so I'm not 100% certain about this), it would be neat if they gave her an append that used Chihiro's real voice* too! That way we could have the original Yukari and a super-realistic Yukari, for fans of both. I can understand why, while AI voicebanks are extremely cool, a character as popular as Yukari would have fans that would miss the original.

*Providing she's comfortable with an exact replica of her voice being made, which is another reason why they might not have used her samples.

 

[Rylitah]

Chihiro Ishiguro literally gave up her career for Yukari; I don't think she'd have a problem recording a faithful AI bank, haha. It does make me wonder why they didn't, since I think Yukari's AI bank just used Vocaloid renders too.

Honestly I thought that a demo of one of the initial announced vocals would finally start answering things, but now I'm just kinda left more confused than ever.

 

[___]

But regarding her silhouette...I saw it and couldn't unsee this