I’m getting a bit tired of ai vbs. I feel like ai vbs tend to sound a bit more similar to each other than sample based vbs, and I just like the sound that sample based vbs produce. But maybe it’s just my subjective opinion
Unpopular Opinions
- Thread starter TheStarPalace
- Start date
That's what I also didn't like, I agree, there was valid criticism but so often it was also hard to engage with it due to it being often stated in bad faith, with no desire for improvement, but to see the thing you don't prefer die and crumble, like for example, there was no reason to be mean about the nana contest, Yamaha giving fans opportunity to voice a vocaloid is probably the most doujiny thing they've done in long time and definitely step in the right direction, it's fun and engaging, I see it as nothing but a positive thing, there was no reason to be mean and nitpicky, yet people were... I understand being saddened that it was targeted at the japanese crowd exclusively but I imagine it'd complicate things and it was easier to realize this way, beside half the people didn't realize it was targeting the eng community.
I have always worried about coming across like I'm treating Yamaha and Techno Speech like my buddies too, but as a newbie in the fandom I was discouraged from getting V5, people were acting like it'd make my computer explode, but when I proceeded to get it with caution, I ended up really liking it, and ever since then, I've lowkey made it my mission to offer alternative perspective on disliked editors in the eng community, in highlighting their strengths, because I know there will be people for whom it's the right choice.
I don't get it either, even with that context, but also, I don't have the same history with those characters like the chinese community does so it's fine for me to not understand.
Speaking for me personally however, I adore those designs, I do see myself using them once I get my hands on their V5s (
).
Tbf that's not the original design they debuted at first. The original designs for the Vsinger V5 voices were just really casual, too casual.Pardon me for derailing the topic, but since v-synth character designs came up I just wanted to say two things:
1. Miku V4x is cute the people who don't like it are just mean
2. I think this one only counts as an unpopular opinion in the Chinese vocaloid fandom, but I cannot understand for the life of my what the heckie the Chinese fandom's problem was with Vsinger V5 designs. Like I read a twitter thread once about how vocaloids in China are much more like virtual idols with established identities as opposed to the mostly blank states with the author of the thread suggesting that might be a factor, but even from that perspective I don't get it. Like what was it about these designs that made the characters feel so radically different it deviated from their established identity? Like... It's just them but in slightly different clothes... I don't get it...
Like... this looks cute???? What is wrong with it?
I recently realised that for me, the most important parts of a cover are the ones talked about the least: layering and processing the vocal tracks in a DAW to achieve the same 'thickness' as the original, and gain staging (setting the volume of the vocals and instrumental so neither is too loud or too quiet). I would genuinely enjoy a cover with neither tuning nor EQ, providing the volume and intensity of the vocals were on point.
Last edited:
Making a vocal synth sound squeaky and/or adding growl does not make the vocal sound more powerful or like it has more emotion.
Despite my love for Avanna, I am not at ALL a Yameii fan (OSEAN WORLD is super weird and uncomfortable to me) and I'm slightly annoyed that that's often what she's known for these days. At least the much more pleasant (to me anyway) notion of Avanna as the "Porter Robinson loid" has come back a bit with Po-uta's release.
TBF, OSEAN probably has every right to use Avanna in that way. I'm sure there are Oliver fans that hate Bee and Puppycat as well. (I don't, but I'm sure there are those that do)
TBF, OSEAN probably has every right to use Avanna in that way. I'm sure there are Oliver fans that hate Bee and Puppycat as well. (I don't, but I'm sure there are those that do)
I've seen many Vocaloid fans disliking Yameii and they had really valid reasons to. Like I did saw accusations of OSEAN being a weirdo and it's okay not to like the concept. Personally I like characters like Yameii, Perty (similar but more hyperpop rap and Cyber Diva is used) etc. simply because it's a creative way to use the loids
I actually like how most of the NT vbs sound, excluding Luka ( poor Luka never catches a break ).
There's other issues to be had, but talking about sound exclusively, I wonder how many people just nitpick for the sake of nitpicking because "Piapro NT bad " is part of the status quo.
With that being said, do I genuinely believe that if people ""opened their hearts"" everyone would suddenly love NT's sound? No lol, definitely not, it's sound is somewhat of an acquired taste and I can see how it would genuinely not be everyone's cup of tea, so this is not disregarding your preference, just wondering how many people are playing into the status quo, I obviously can't read minds so we'll never be able to tell.
Often I hear complaints and I just don't hear it/the issue, but this type of stuff is highly subjective so *shrugs*
There's other issues to be had, but talking about sound exclusively, I wonder how many people just nitpick for the sake of nitpicking because "Piapro NT bad " is part of the status quo.
With that being said, do I genuinely believe that if people ""opened their hearts"" everyone would suddenly love NT's sound? No lol, definitely not, it's sound is somewhat of an acquired taste and I can see how it would genuinely not be everyone's cup of tea, so this is not disregarding your preference, just wondering how many people are playing into the status quo, I obviously can't read minds so we'll never be able to tell.
Often I hear complaints and I just don't hear it/the issue, but this type of stuff is highly subjective so *shrugs*
While on the topic of that, I can't stand how people are saying that Miku is becoming irrelevant due not having an AI
Who's the one still getting video games, merch, and concerts? Still a virtual celebrity? Miku's not going anywhere.
Besides, you hear any vbs newer than V4 in that song for the new Guardians of the Galaxy? It's very likely the V4 editor was used, even moreso if Piko was really among those voices.
I also dislike that Talkloids are portraying Miku as bitter over everyone getting AI vbs. I'd like to see a take that she's not pursuing an AI voicebank, she's sticking it to the man and staying a robot.
Who's the one still getting video games, merch, and concerts? Still a virtual celebrity? Miku's not going anywhere.
Besides, you hear any vbs newer than V4 in that song for the new Guardians of the Galaxy? It's very likely the V4 editor was used, even moreso if Piko was really among those voices.
I also dislike that Talkloids are portraying Miku as bitter over everyone getting AI vbs. I'd like to see a take that she's not pursuing an AI voicebank, she's sticking it to the man and staying a robot.
I was telling myself the same thing, people portraying Miku as bitter over other synths is kinda funny to me, as if Miku wasn't and isn't the one holding the monopoly over all other voicebanks, even Kafu, who's the 2nd most popular synth after her, doesn't get even half the amount of content made featuring her as Miku ( saying it figuratively but wouldn't be surprised if it were statistically true too ), I find that us, overseas fans greatly underestimate Miku's impact.
Beside that, I find that content to be massively childish, it's like the 2020s version of Rin or Gumi being jealous of Miku's popularity even tho "they're so much better"
I don't know anything about the artist, I just don't really like the style at all. AVANNA's also known for Porter Robinson though so there's that as a more positive association for me.
Yeah, I see it that way too. I don't see her feeling the need for an AI VB at all really. Teto's AI sounds just like her (usually; a couple of her Vocal Modes seem to stray from the classic tone some), but I think that's partly because her voice provider can actually sing in her Teto voice; I've heard Saki Fujita's Miku voice is such a unique timbre for her that she can't sing in it very well. SVC like the unreleased Cherry Pie could work, but making Miku AI work for SVS might take more work than it's worth.
tbh i don't really get people who r like "omg every vb should get an ai update!". Sure some vbs that are discontinued or are of poor quality would benefit from an ai vb. But for the most part most vocaloid vbs are either are of good or some even of very good quality, so i don't really see a big point in them getting an ai vb other than maybe being able to sing in 4 languages if they go to sv and voalmodes and stuff. Like for example, cryptonloids do not need ai updates for the most part, they have a lot of appends and vbs, everyone has an eng vb and miku even has a chn one, and you can make further use of them with piapro studio.
pls don't jump on me im just saying my feelings TwT
pls don't jump on me im just saying my feelings TwT
Recently I saw a discussion on Twitter between some users about Ci Flower. "I don't understand," they said. "Sometimes, Ci flower sounds exactly like the flower I know, expressiveness and all. And other times, she just sounds like 'a CeVIO'".
I thought this was interesting because I have rarely ever seen a discussion 'in the wild' with this nuance. The "sound" that people associate with v Flower now is one that developed over the course of years and people pushing the engine to extremes. People can do that with CeVIO, too-- but often don't.
This is all to say that there is still whole lot of potential for genuine artistry through the use of AI voicebanks, but we might not have gotten to the point of that being the "standard" of use yet. Nowadays, we expect basically unique sounds/usage/tuning from all of our favorite producers using VOCALOID. But that hasn't quite crossed over to the world of AI yet. I think there was a lot less variance in sound on VOCALOID in the early days of the engine, but it started to change when knowledge and possibilities started to disseminate over the years.
At least on Synthesizer V, a lot of resources have been poured into automation features to make the engine as beginner-friendly as possible, which makes sense as the learning curve was a huge issue for adoption in the professional market for VOCALOID. But we don't really have that many SynthV power users in the realm of original songs. Just like the beloved engines of old, AI-based engines are still instruments that can be "played" the way you want to hear them. It's just a different instrument from what we had before, and the initial adoption is still ongoing. On CeVIO AI, there are so many examples now of stand-out, differentiating, and unique usage, but they are still uncommon in the overall pool.
And with that in mind, I don't think there's anything wrong with users wanting to see AI versions of their favorite voices. But the mindset of it being a progression of linear upgrades: standard -> AI is wrong. Why not offer high-quality, sample-based voicebanks for a character and an AI model? There are unique advantages in both systems and they don't completely occupy the niche of the other. It bothers me that so many fans have a mindset on either extreme: "AI bad" or "AI's the best, everything else is outdated", that ultimately limits the creative expression of these technologies. We've also seen fascinating combinations of both following the release of VOCALOID 6, so the tribalism has become even more confusing to me.
WRT Hatsune Miku specifically though, I think Crypton has been correct about their steadfastness not to give Hatsune Miku an AI voicebank given what we know about the development of the "Hatsune Miku" product specifically and how important VOCALOID and NT's format is to her sound and character.
I do want to see something unique/exciting happen with her technology, though, and I can understand the perspective that the absence of excitement surrounding Miku's new technology might not spell well for the long-term, even if V4, games, and merchandising are more than large enough a wave to ride on. But I am at a loss as to what that is, and I can't imagine the brainstorming has been any easier for Crypton. If there were an obvious "good option", I'm sure it would have happened already as the explosion of the VOCALOID renaissance would have been the perfect time for a new release, from a commercial perspective.
EDIT: speaking of all of this, I'm really glad Kasane Teto is the "core character" that has both extremes at once first-- a beloved series of classic VBs and a wildly hyped AI VB. Her platform was already one of experimentation and fan energy rather than prestige. Depending on how this goes, we might see a whole lot of new stuff come out, or not much will change at all.
I thought this was interesting because I have rarely ever seen a discussion 'in the wild' with this nuance. The "sound" that people associate with v Flower now is one that developed over the course of years and people pushing the engine to extremes. People can do that with CeVIO, too-- but often don't.
This is all to say that there is still whole lot of potential for genuine artistry through the use of AI voicebanks, but we might not have gotten to the point of that being the "standard" of use yet. Nowadays, we expect basically unique sounds/usage/tuning from all of our favorite producers using VOCALOID. But that hasn't quite crossed over to the world of AI yet. I think there was a lot less variance in sound on VOCALOID in the early days of the engine, but it started to change when knowledge and possibilities started to disseminate over the years.
At least on Synthesizer V, a lot of resources have been poured into automation features to make the engine as beginner-friendly as possible, which makes sense as the learning curve was a huge issue for adoption in the professional market for VOCALOID. But we don't really have that many SynthV power users in the realm of original songs. Just like the beloved engines of old, AI-based engines are still instruments that can be "played" the way you want to hear them. It's just a different instrument from what we had before, and the initial adoption is still ongoing. On CeVIO AI, there are so many examples now of stand-out, differentiating, and unique usage, but they are still uncommon in the overall pool.
And with that in mind, I don't think there's anything wrong with users wanting to see AI versions of their favorite voices. But the mindset of it being a progression of linear upgrades: standard -> AI is wrong. Why not offer high-quality, sample-based voicebanks for a character and an AI model? There are unique advantages in both systems and they don't completely occupy the niche of the other. It bothers me that so many fans have a mindset on either extreme: "AI bad" or "AI's the best, everything else is outdated", that ultimately limits the creative expression of these technologies. We've also seen fascinating combinations of both following the release of VOCALOID 6, so the tribalism has become even more confusing to me.
WRT Hatsune Miku specifically though, I think Crypton has been correct about their steadfastness not to give Hatsune Miku an AI voicebank given what we know about the development of the "Hatsune Miku" product specifically and how important VOCALOID and NT's format is to her sound and character.
I do want to see something unique/exciting happen with her technology, though, and I can understand the perspective that the absence of excitement surrounding Miku's new technology might not spell well for the long-term, even if V4, games, and merchandising are more than large enough a wave to ride on. But I am at a loss as to what that is, and I can't imagine the brainstorming has been any easier for Crypton. If there were an obvious "good option", I'm sure it would have happened already as the explosion of the VOCALOID renaissance would have been the perfect time for a new release, from a commercial perspective.
EDIT: speaking of all of this, I'm really glad Kasane Teto is the "core character" that has both extremes at once first-- a beloved series of classic VBs and a wildly hyped AI VB. Her platform was already one of experimentation and fan energy rather than prestige. Depending on how this goes, we might see a whole lot of new stuff come out, or not much will change at all.
Last edited:
I do wish that more of the AI vbs also had a standard type vb. (I believe that Tsuina-chan was the last Synth V to have a standard vb as well as an AI vb). Trying to use an AI vb alongside any of the standard types can be pretty jarring to me and ends up hampering my desire to aquire more of them.
I’ve been debating putting Rime and Solaria on my wanted vb list, but my thoughts keep coming back to when I’ve tried using Saki’s AI alongside the loids I already owned and having the difference in texture actually set off my ocd of all things.(I got Saki AI through the Saki standard promotion) I was very pleased when Mai was released with Synth V pro, because now I have a duet partner for Saki without having to purchase more vbs.
The last SynthV to have a standard vb was Kyomachi Seika, but she was released the same month as Solaria (who ironically also once had a standard vb in earlier stages of development)
I think SynthesizerV might just not be the best platform for "standard" VBs at this point in time-- and to be honest, it never really was. It was kind of different back in the early days when the distinction to std/ai didn't exist yet or was new to the engine, and people were really thinking about the potential of QOL and feature additions for STD specifically and that contributed to the excitement, but what focus that remained in that area is not there anymore. At least for Japanese, and with only the resulting sound in mind, STD Synthesizer V was still kind of struggling in quite a few areas despite its good additions.
It'd be cool to see voices exist simultaneously across multiple engines, but commercially VOCALOID is really far and away still the very best possible option for "standard" synthesis, and who knows what "making a voicebank for VOCALOID" looks like now. The nana AI VB still isn't out, Po-uta was Yamaha... the only third party is AI Megpoid. Maybe the licensing is just still expensive?
It'd be cool to see voices exist simultaneously across multiple engines, but commercially VOCALOID is really far and away still the very best possible option for "standard" synthesis, and who knows what "making a voicebank for VOCALOID" looks like now. The nana AI VB still isn't out, Po-uta was Yamaha... the only third party is AI Megpoid. Maybe the licensing is just still expensive?
imo SVS is pretty weak engine in terms of concatenative synthesis but I would have still loved for that to have been pushed to the extreme:
Considering 3 of the standard vbs on the engine are literal UTAU ports, I'm pretty certain that the standard vbs are basically CVVC UTAUs in practiceabout Renri...she's VCV but you can oto CVVC aliases for VCVs ) and getting just plain 4 pitch CVVC vb in the most vanilla form...isn't super exciting, it would've been really exciting for them to do multi-append vb, integration of such vb would've been especially smooth considering the engine has it's own equivalent of suffix broker = expression group.
Concatenative synthesis went really unrealized on the engine
Considering 3 of the standard vbs on the engine are literal UTAU ports, I'm pretty certain that the standard vbs are basically CVVC UTAUs in practice
Concatenative synthesis went really unrealized on the engine
Yeah. SynthV Standard voices are often clear but many of them are muddy, with wonky pronunciation (especially in English) and an odd "beep-y" undertone. I much prefer the sound of Vocaloids when it comes to standard synthesis myself. SynthV AI sounds GREAT though. I think the introduction of AI VASTLY improved SynthV's sound.
That's true. Would've been nice for the concatenative VBs to have something stand-out, y'know? I feel like for concatenative synthesis SynthV R1 was more fun
I've always agreed with the sentiment, but I could never put my finger on it, the most stand-out standard vbs do after all come from R1, majority of which were chinese, and to my ( untrained ) ear chinese concatenative synthesis "felt" the best to me on SynthV.
Kinda like how DV is optimized for chinese and chinese synths have historically performed the best on it while japanese always felt off to me, and that's a feeling I share about standard SynthVs too. Like as an example, as unique as Genbu's tone is...he's super wonky on the engine and I do find GAO superior to him.
It's hard for me to speak on english, it's hard to make assessments when there are only two standard engloids.
I don't speak japanese or chinese so I'm purely talking out of my ass, don't put too much weight behind my words.
Oh yeah! I remember Japanese DVs sounding weirdly robotic, almost reminiscent of old vocal synths, while Chinese DVs have sounded smooth from what I've heard (I've never used one though due to not speaking Chinese nor knowing a plethora of good Chinese songs).
Maki English Standard might be my least favorite English voicebank. Her pronunciation is extremely wonky and her accent just feels "hard" to me. Her AI voicebank makes her accent softer and more pleasant to the ear. And yeah, I agree on Genbu, the SynthV engine does not do him justice, and on that note, UTAU Renri sounds WAY better than SynthV Renri IMO. She's brighter, clearer and more expressive on UTAU for sure.
SynthV Standard banks have an overall "harder and sharper" feel compared to Vocaloids and I feel like they're more likely to sound kinda jarring for it? Even the Vocaloids that really enunciate, such as Miku V4X and (especially) AVANNA, don't sound quite that way. I feel like they put way too much into making them clear and strong in attack and it gives them this really weird sound?