Unpopular Opinions

[Scarlet Illusion]

Yeah, I feel you with AI banks, though, I think in Kafu’s (and other CeVio AI’s) case it’s an issue of limited editing options and also you can’t turn off the AI tuning. (I haven’t used it myself, but my understanding is that if you want to “tune” them how you would any other synth.
I like some SynthV AI’s (like Rikka and Maki English) but I find that I can’t use the AI tuning (which is optional, luckily) or else it feels like there’s not much that can be altered? Or maybe that’s moreso because I’m not comfortable tuning in SynthV the way I am in Vocaloid oR UTAU… but if I’m not using the AI tuning features, then what’s the purpose of having the bank? I feel like there’s not much of a middle ground or room for people to tune with them.
Though, I have heard a few nice uses of them, such as the song Peaches mentioned and Police Piccadilly’s demo for Kafu, which still remains the sole usage of Kafu I’ve heard and liked.

 

[peaches2217]

*cracks knuckles*

Kaito V3 English might just be Crypton's best English bank. Subjectively speaking, he has a very pleasant and soft tone, but isn't weak; objectively speaking, he's smooth and transitions between words well. The only phoneme editing I've done with him is to have him enunciate words in a certain way. I haven't had to touch his phonemes for accuracy-related purposes. I think he easily surpasses all but maybe Miku V4X English, and even then I prefer him to her. He's quickly becoming my favorite non-native English vocal from any company on any engine the more I work with him.

 

[blackout]

Lemme ramble for a bit... as much as I like SynthV and think it's killing it as an editor I really don't want to see it become the One and Only commercial singing synths which seems to be what a lot of ppl seem to want to happen.

Like I get it, especially after its latest update I can def see why you would want your fave voice to be on it but I don't want everything else to die out and move over just because of that. It's honestly a bit obnoxious to see so many ppl declare it as a Cevio or Vocaloid killer when it's not even close when it comes to recognition (or sales..) to either of those.

Miku NT for instance, she's still pretty flawed as a product but I don't want them to give up on NT entirely. I really dig it as a concept and frankly I do think they're genuinely trying and improving it with every update. Don't get me wrong though, I don't think they should've passed it as a finished product for the price its sold and kept it as an Early Access kinda thing marketing wise but nevertheless I think the future is looking brighter for them in this front and I'm honestly excited to hear how she'll sound as they continue developing the engine.

I don't believe it'll become a flat out replacement for Vocaloid either... Let's face it just on brand name alone Vocaloid will always be king. The only non-Vocaloid commercial singing synth that's doing as well as the currently popular Vocaloids is Kafu on Cevio AI. Besides this we're also experiencing another peak when it comes to Vocaloid music that we hadn't seen since 2012-2013. Now I know Yamaha is kinda... Unpredictable and V5 has been a weird era of questionable decisions (the fact there's still no cheaper option of the V5 editor that doesn't come with the default voices is beyond me) but I'd like to HOPE that they've acknowledged this and the silence just means they've been cooking up something else all along... I honestly just want to see Vocaloid improve.

Idk I just think that the monopoly Vocaloid used to have on commercial synths was what made it stagnate as an engine after V3 and I'd hate to see that happen again. Competition is healthy and we wouldn't have SynthV be this good if Dreamtonics hadn't looked at the current market and went "hmm how can we make our product stand out". I'm just glad we have so many more options today than we used to, I think every editor has its pros and cons and certain things they do better than others and I hope Dreamtonics, Techno-Speech, Yamaha, Crypton & etc continue to go their own ways and improve on what they have rn.

 

[Prism]

I think if the budget allows I think vocal characters can be and should be on every synth they offer different pro and cons. Take synth v especially with the cross lingual ability they can offer their vocal characters in different international markets. I could see zunko and co getting a synth v because of it because they have been wanting to offer zunko and co in English and that would justify making another bank for them. Looks at Crypton future media this is really what they should try to do in my opinion they have the money project Sekai made 5 Millon last month on just on Android. I understand they are going for a more robotic stylized voice (which doesn't really make sense in my opinion because producers like Mitchi m found massive success because of their realistic tuning to the point where they are working with Sega and crypton and has changed the image of their voice ). I'm not saying their goals with nt isn't warranted, I'm just saying they should offer their characters on more because different synths offer different things that appeal to different people it's just good for business.

 

[pico]

I think if the budget allows I think vocal characters can be and should be on every synth they offer different pro and cons.

Yeah, this is the dream.
Like, when I say “I wish I could see more of my favorite vocals on X new engine or system” it doesn’t necessitate the removal of competition or variety in the synthesis space.

That said, I can’t say I see the Crypton characters ever existing on a system outside of Piapro for the time being because they need people to buy into their engine. It doesn’t seem like the best business move for them at this moment considering they know NT is in early stages and need people to stick with the engine to prevent it from being completely washed away by their competition— the voices are, at this moment, their leverage.

But for vocals from companies without the necessity of marketing their own synth technology, it’s a fun dream to envision a future where they might exist on many engines and afford the user more choice and flexibility, like Kiritan.

 

[Pewe]

I think the “Crypton is aiming for robotic voices” is a bit of a misconception and not *entirely* true. If Amano Kei’s analysis on Miku NT and Piapro NT is correct, the main reason Crypton wants to stick to the “robotic ness” is bc it’s part of the Miku charm, not because they want to make her vocals as choppy as possible (which is why the choppiness and some unnatural ness have progressively been fixed with the updates).

Seeing people always hating on NT even going as far as comparing Miku NT to a cheap CV UTAU (which honestly is FAR from the truth) always leaves a sour taste in my mouth bc if, again, Amano’s analysis on how NT’s processes the voices is correct, then what Crypton is trying (and doing) could be HUGE. It would be the most “pure” version of Miku and Co. we could get, and I *really* doubt that goal could be achieved with Synth V, CeVio or even VOCALOID itself bc they aim to replicate *real* voices (and real not only means “smooth”, it also means natural in most cases. Plus, Crypton has always liked trying to innovate when it comes to vocal synths (whether they succeed or fail), so going from yamaha’s limitations to Dreamtonic’s would probably not be in their best interest.

Lastly, I HATE seeing to many westerner fans nagging and borderline harassing wat on Twitter when it comes to Miku NT. Like c’mon you can hate the program all you want but it’s not the man’s fault. Even if you bought her and don’t like her, the overpriced price is on the executives, he’s just the developer.

 

[Prism]

I only said crypton because they have wanted and struggled in the past offering their characters in English and Chinese. I don't see them adding non Japanese language support to nt any time soon. I don't want to come off as hating on nt, starting a new synth is a very difficult especially with COVID.

 

[Pewe]

I only said crypton because they have wanted and struggled in the past offering their characters in English and Chinese. I don't see them adding non Japanese language support to nt any time soon. I don't want to come off as hating on nt, starting a new synth is a very difficult especially with COVID.

Oh no, I didn’t mean you lol. I was talking about people on Twitter that harass wat and others that for some reason feel the need to express their NT hate almost everywhere

 

[peaches2217]

I think the “Crypton is aiming for robotic voices” is a bit of a misconception and not *entirely* true. If Amano Kei’s analysis on Miku NT and Piapro NT is correct, the main reason Crypton wants to stick to the “robotic ness” is bc it’s part of the Miku charm, not because they want to make her vocals as choppy as possible (which is why the choppiness and some unnatural ness have progressively been fixed with the updates).

Seeing people always hating on NT even going as far as comparing Miku NT to a cheap CV UTAU (which honestly is FAR from the truth) always leaves a sour taste in my mouth bc if, again, Amano’s analysis on how NT’s processes the voices is correct, then what Crypton is trying (and doing) could be HUGE. It would be the most “pure” version of Miku and Co. we could get, and I *really* doubt that goal could be achieved with Synth V, CeVio or even VOCALOID itself bc they aim to replicate *real* voices (and real not only means “smooth”, it also means natural in most cases. Plus, Crypton has always liked trying to innovate when it comes to vocal synths (whether they succeed or fail), so going from yamaha’s limitations to Dreamtonic’s would probably not be in their best interest.

Lastly, I HATE seeing to many westerner fans nagging and borderline harassing wat on Twitter when it comes to Miku NT. Like c’mon you can hate the program all you want but it’s not the man’s fault. Even if you bought her and don’t like her, the overpriced price is on the executives, he’s just the developer.

 

[Prism]

Oh no, I didn’t mean you lol. I was talking about people on Twitter that harass wat and others that some why feel the need to express their NT hare almost everywhere

I think that's because a lot of people are not happy with NT. I've seen a lot of people in Japanese and Western communities voice their problems with it and if sales are any sign of it, it's very telling people don't like or want NT. I think a lot of people with the pre sales and were burned when they were given a buggy beta like software and how crypton handled it. We all love vocal synths and have an emotional attachment to to this software so it's understable for people that were burnt to express their issues with the software but it doesn't excuses harassment.

 

[Pewe]

Oh yeah I totally get it (even I, who likes current Miku NT, HATE how they released it and how Luka and Meiko sound), but some people really go out their way to put wat and his project down even in his mentions which is kinda....

 

[peaches2217]

I think that's because a lot of people are not happy with NT. I've seen a lot of people in Japanese and Western communities voice their problems with it and if sales are any sign of it, it's very telling people don't like or want NT. I think a lot of people with the pre sales and were burned when they were given a buggy beta like software and how crypton handled it. We all love vocal synths and have an emotional attachment to to this software so it's understable for people that were burnt to express their issues with the software but it doesn't excuses harassment.

There's been a lot of critiques, but financially she's been been doing well so far. Her sales don't rival Miku's previous VBs, but she's in the Top 3 on Sonicwire pretty regularly (and is at #1 right now).



It's not that people don't want NT. Plenty of people wanted and continue to want her, but there's critiques about her quality and how Crypton has been handling her, all valid. God knows I have a thing or six to complain about! That girl is BUGGY. It's kind of like people's assumptions about V5: lots of people are having issues with her and don't care for her, so they say she's a financial and commercial failure, and that's just not true.

And, as Pewe said, Wat's getting dogpiled for things that were never his fault to begin with. It doesn't matter what you think of Crypton or NT. That behavior is just disgusting, and it needs to stop. (Thankfully I haven't seen anyone here doing it! But my heart goes out to him. He deserves so much better.)

 

[Prism]

There's been a lot of critiques, but financially she's been been doing well so far. Her sales don't rival Miku's previous VBs, but she's in the Top 3 on Sonicwire pretty regularly (and is at #1 right now).

View attachment 5930

It's not that people don't want NT. Plenty of people wanted and continue to want her, but there's critiques about her quality and how Crypton has been handling her, all valid. God knows I have a thing or six to complain about! That girl is BUGGY. It's kind of like people's assumptions about V5: lots of people are having issues with her and don't care for her, so they say she's a financial and commercial failure, and that's just not true.

And, as Pewe said, Wat's getting dogpiled for things that were never his fault to begin with. It doesn't matter what you think of Crypton or NT. That behavior is just disgusting, and it needs to stop. (Thankfully I haven't seen anyone here doing it! But my heart goes out to him. He deserves so much better.)

Was mainly looking at Amazon Jp to look at her in the context of whole industry she tends to fluctuate between 20s in box and 50s for downloads but I do imagine she would do better on sonicwire. What i did find interesting about the rankings is that Rin and Len v2 sells better than their V4 at least right now at the moment.

 

[choripan]

my new unpopular opinion is that i dont want miku/rin/len/luka ai. they were made to have that sample based-synth-noise as a charming aspect, and they will thrive on a proper nt engine... which rn... it's not. meiko and kaito? yeah those were made with being "professional" in mind, and i feel like crypton has been kinda butchering them...

 

[Blue Of Mind]

my new unpopular opinion is that i dont want miku/rin/len/luka ai. they were made to have that sample based-synth-noise as a charming aspect, and they will thrive on a proper nt engine.

I'd say Meiko, Kaito and Luka would fit the best for AI since they have more "realistic" sounding VBs, rather than Miku, Rin and Len who are all clearly voice acted. Saying that, I wouldn't be opposed to AI versions of Miku, Rin and Len either, simply because I'd like to hear the end results of any AI banks for them.

Once Piapro NT gets out of its perpetual beta state, and ideally adds English and Chinese language support first, it would be cool to see an AI update for it.

 

[Prism]

This is just a look into what Crypton characters would be like if they got an ai bank. This is just mix of opinion and speculation and a show case of their voice providers singing.
I also think crypton threw Meiko and Kaito into the CV series to their detriment. I would love to hear what an ai Meiko would sound like her voice providers Meiko Haigou naturally has a beautiful soft airy voice that is so different from vocaloid Meiko.

Kaito sounds closer to his vocaloid probably because the advancements in tech. Ai would suit his voice really well for a ai Kaito.

Now with Miku she's very iconic and has a hard time being replicated. It's probably why they keep using her v2 samples. I wonder if they can train a voice bank on those samples like Ritsu enunu or the medium 5 ai banks. I think saki fujita singing voice is pretty and would make for a beautiful bank you can hear Miku coming out occasionally. Saki Fujita is a master voice actor so she could probably get her voice closer to Mikus.

Now Rin and Len would be a perfect fit for ai. Tsuina Chan has shown us that we can have a good heavily voiced acted ai vocal synth and Asami can do the voice. Granted the ai banks might be softer and doesn't have that robotic power they are known for like Meiko. They would probably sound closer to their warm append.

Now Luka, Yu Asakawa in past interviews has said it was hard to voice act and sing the vocaloid recording script. So an ai bank would solve that problem while introducing a new one. Yu Asakawa has also said in the past she's afraid of vocaloid taking her job. This is probably why we haven't seen cyrpton tts yet. I think ai tech would help Luka the most with being able to comfortably voice act I could see a new level of expression without having to deal with her wonkyness found in her V4 release.

Cryptons Characters have a certain charm to them that I think can be persevered into ai. Ai can also open up certain genres of music that vocaloid has struggled in the past in like ballads and acoustic. This wasn't to bring their efforts in nt down. I think nt and ai are both valuable and have their place in the vocal synth landscape.

 

[lIlI]

I love listening to and admiring the work of expert tuners; however when it comes to voicebank demos, I prefer to hear how the product sounds with minimal editing. If a vocal synth needs a lot of work to sound impressive, I want to know that before I invest in it. A demo by someone who can make any voicebank sound good may unintentionally mislead. It's good to know a voicebank's potential, but knowing the baseline you'll have to start from is vital too.

The most useful voicebank examples are for me, the wave of plug-and-plays that come immediately after a product releases. Genres that don't fit, tempos that are much too fast, little tuning: this is legitimately the most revealing and informative period for me as a buyer. It shows the limits of what the product can do, and what will need extra elbow grease to work.

I understand why the prevailing idea in the fandom is 'the best tuners should make demos', and I don't expect a company to intentionally show a product's limitations. However, for practical reasons I most appreciate a demonstration that shows what a beginner could achieve!

 

[Leon]

I like this take! I think I'd like a mix of both (little to no work vs a good bit of work, just to hear the base and also some of the best results possible), but you make very solid points.

 

[Prism]

I wish companies would make tutorials showing you go from raw to tuned to completely show you what your getting and what it takes to get there.

 

[ikaroll]

I think the demos from the company/studio themselves should have minimal tuning, and then leave the complex tuning to the producers who make their own demos. That way, you can see what the bank sounds like by itself, and then people can enjoy how the other producers tune their songs!