Unpopular Opinions

[___]

Maybe it's just me being a little out of touch with the UTAU side, but I was under the impression that Defoko was universally appreciated across the fandom? Most people seem to recognize that she's a fully synthetic voice and like that about her.

Oh she definitely is! And she def has her appeal but also her being completely synthetic and ( probably at the time due to the limitations of tts ) incapable of producing longer samples doesn't ( ironically ) make her very standard and good representation of what the engine can do. She's def there with Chipspeech which has it's place among voice synths but it's not necessarily representation of the sound UTAU can produce.

 

[Rolo]

I honestly have to agree. She’s spectacular! I don’t voice my opinion on her much but her voice is one of my favorites, child loid-wise.

Interestingly enough Rana is actually a high school girl and might be closer to 17-18 according to this post from the official account which lists her as a 3rd year high schooler

 

[peaches2217]

Interestingly enough Rana is actually a high school girl and might be closer to 17-18 according to this post from the official account which lists her as a 3rd year high schooler

Ah, that’s embarrassing! I always headcanoned her as a child. Goes to show how much I know. Point still stands: she’s criminally underrated.

 

[Prism]

People have made a vcv defoko and she sounds great.

 

[Rylitah]

Ah, that’s embarrassing! I always headcanoned her as a child. Goes to show how much I know. Point still stands: she’s criminally underrated.

To be fair, pretty much everyone sees her as a child, haha. Probably has to do with her "canon" age of 0, but apparently that was actually updated each year on her Vocaloid anniversary, which is what it was really counting, but.... you know, since she's unpopular, it's not like people really followed her to know that.

Rana's one of those Vocaloids that I'm always on the fence about buying and I get really intense urges to but for some reason never have (probably because I heavily prefer physical boxes over downloads). Maybe someday she'll end up in my collection...

 

[Prism]

I hope one day someone translates the magazine into English become the rana concept is great

 

[Twillby]

Speaking of Vocaloids and ages, am I the only one who didn’t think Iroha was obviously a child? I knew a lot of really skinny and short young women in college (like I could lift them in the air and I’m weak as heck) so maybe that’s just my personal history interfering with my perception. Feels super embarrassing to admit now that I know Kyounosuke might have been around 15 when he recorded her V2, but yeah.

 

[Leon]

...I always saw Iroha as an adult, oof. Then again, I tend to disregard Vocaloid "canon" ages entirely anyhow, and adjust them per son as needed. Vocaloids are too fluid to focus really hard on the ages imo, esp when a lot of them aren't given "official" ages anyhow :>

(On another note, could someone explain to me how Defoko's samples are entirely synthetic? I've heard that before, but I'm curious as to how exactly that works!)

 

[Rylitah]

(On another note, could someone explain to me how Defoko's samples are entirely synthetic? I've heard that before, but I'm curious as to how exactly that works!)

Iirc her samples are from the tts Aquestalk (the tts Pinocchio-P uses in his songs), which is why they're so short, since they're just chopped up speech samples. You can actually mess with her on the online demo here, she's Female-01 (f1 in the dropdown list).

I assume that this tts is computer generated, though I don't know too much more about the Aquestalk side of things.

 

[lIlI]

In my opinion, age misconceptions have led to some of the funniest arguments in the fandom, my favourite being everything based on 'Fukase is canonically 30', because Fukase is the babiest looking Vocaloid since Ryuto. Try reading any Youtube comment based on that rumour while looking at this face.

Me when people say that Rana is a newborn and Fukase does taxes

 

[DefiantKitsune]

your daily reminder that Rana can book her own doctor appointments in most of the US but Qingxian cannot

as someone who doesn't care for Defoko, she seems pretty popular- one of the more commonly used vocaloids for originals with Teto and Renri and maybe Ritsu in the last few years

 

[cafenurse]

Iirc her samples are from the tts Aquestalk (the tts Pinocchio-P uses in his songs)

oh i always thought he just used defoko lmao. the more you know!

 

[Blue Of Mind]

This discussion on ages made me remember when I thought both Kaito and Meiko were officially 21 because of random fanon I saw on YouTube... (Officially they have no canon ages, but the fandom tends to depict them as adults, perhaps a bit older than Luka who's officially 20).

I think the only canon ages people take notice of are the Cryptonloids' ages. Many people depict Miku as 16 and Rin and Len as 14, as stated in their bios. Luka is sometimes all over the place, I've noticed, so I think people are less aware that she's officially 20.

 

[___]

Saki Standard > Saki AI

I don't like how much power she lost in her AI compared to Standard and her vocal range noticeably shrunk.

She may be more realistic and not produce as much engine noise but her output fails to produce varied sounds and on more subjective note she sounds really bored out of her mind and generally is an unispired vb for me.

 

[peaches2217]

Saki Standard > Saki AI

I don't like how much power she lost in her AI compared to Standard had and her vocal range noticeably shrunk.

She may be more realistic and not produce as much engine noise but her output fails to produce varied sounds and on more subjective note she sounds really bored out of her mind and generally is an unispired vb for me.

Having run tests on her, I have to agree. She's a fine VB, and she lacks the hiss and engine noise that her Standard has, but her Standard just feels so much more unique to me. Her AI bank impressed me for a day or two and then I quickly got tired of it.

 

[Prism]

I really think both of them have their own merits. I do see saki regular to be a more throat singing and ai to be more on the nose side ak brighter I wonder if you can play with the tension to fix it. I do enjoy ai better, I think it captures more of her vp tone. But I can see were people enjoy the characteristics of sample based. I'm glad synth v is releasing both. I Am impressed at saki ai's small file size.

 

[___]

Ah I forgot to reply! It's not necessarily abt AI vs sample-based but just how she came out sounding to me, she's so far SynthV's only AI bank so I'm saving my judgement regarding their AI.

 

[KingPapillon]

From what I've found playing with Saki vs Saki AI, is that regular Saki sounds way more synthetic as a default without lowering or tweaking any parameters at first, while Saki AI sounds kind of rough if you use the same kind and amount of parameter tweaks. With default Saki (and all non-ai synth v banks I suppose), you sort of have to "shave" the synthetic throatiness away by lowering the tension on certain phrases, and making use of the other breath and gender parameters, but in return, you can do so much more with her regular voice. You can make her tense, whispery, robotic, or cute, while Saki AI has troubles covering these singing styles as well.

(not to plug my own stuff but)
I made this cover to kind of demonstrate how regular Saki can sound like if you really edit her tension heavily. I think she sounds like a good halfway point between her default unedited voice and her ai voice.. The way I see her two voicebanks, is Saki regular is like having a big piece of marble you can sculpt into many things yourself, and Saki AI is the human looking statue that's already been premade and doesn't leave much for any other interpretation.
I went back to this SVP file after Saki AI came out, and plugged her into this just to see, and yeah, she actually sounded kind of terrible with the edits made for regular Saki..

Overall I'm glad we have both though, love Saki.. And who knows what the future holds, this is our very first AI voicebank released for SynthV after all..

 

[Rolo]

I don’t know if this is unpopular or just more on the controversial side but I would love for crypton to one day just become a lot more daring with Miku’s image and do not only darker songs but also like songs that are just not really what you expect to be performed. Like Kikuo, Maretu, Koyori, Hachi (he’s got some dark stuff) for the more darker side of vocaloid and then songs like Secret Himitsu by Pinocchio-P or even songs by producers like Mer, Monaca:Factory, some of Hachi’s unknown works, powapowa, would be great.

They’re getting better with set lists and adding more interesting songs to the queue but there’s still an over abundance of 2008-2010 sounding Miku songs that I could just do without because a lot are sleepers in my opinion. And I do understand why they do what they do trust me I’ve heard it before, but Miku and by extension the piaproloids are waaay better when the songs chosen for them to perform or that are picked in game are just of a more interesting variety. Not every song has to be dark to be interesting there are plenty of vocaloid songs that cover all types of subject matters but I am tired of like the same 5 Miku songs about her being an idol or helping everyone with her music it’s not 2010 I think we can move forward a bit.

On that note shoutout to project Sekai for having some of the most interesting setlists for a vocaloid game so far.

 

[TheLemonKing]

- I didn't like Sand Planet all that much

- Mafumafu's cover of Snobbism is bad

- I actually don't mind Gumi's v3 design (You know the one with the vest and baggy shorts)

- Party x Party > Bad End Night

- Kenneth's songs were the worst songs in the Communications series

I'm not sure if some of these are unpopular but those are what I could think of off the top of my head.