Unpopular Opinions

[Grzesiek11]

The Japanese fandom and the English fandom don't have to be the same. Japanese people use vocalo (NOT Vocaloid) to mean vocalsynth, so they *are* making a distinction. We have an English word for vocalsynth, we can use it.

Here's an unpopular opinion. "Vocalsynth" is a lame word. Lowercase "vocaloid" is much better sounding, and genericized trademarks are nothing new. Yamaha also holds the trademark to「ボカロ」btw so this is not at all different to using the word "vocaloid" in English, it's just more common to shorten words in this way in Japanese. I wouldn't be opposed to using the word "vocalo" either, it sounds pretty sweet as well. Anything but "vocalsynth".

I'm convinced the word "vocalsynth" was invented by the Big English Speaking Community as a way to continue selling Teto Is Not A Vocaloid to new fans.

 

[Heruru Meruru]

I just call Vocal Synths "'loids" at this point.

 

[IO+]

Thinking about it, I actually prefer using more (not entirely) genre-specific voicebanks to more general tone voicebanks. With genre-specific ones, I can easily know if a voicebank will sound good with the song, I don't have to test anything regarding tone. But general ones? I can't really find a use for them! My best examples are Meiko vs Luka (v4x) English. Luka English has a general pop-adjacent tone, while Meiko is clearly geared towards light pop and ballads. When I look for songs for Meiko, I can easily think of songs that fit her niche, but with Luka? I have no idea! She, along with a lot of other voicebanks, is so general that I simply have no idea what to use her for! Why would I use a general voice that'll get the job done when I can use a voice that fits this niche? I might be rambling, so please tell me if this is incomprehensible and I will do my best to explain.

Vocaloid English are quite troublesome because how their unintuitive phonetic structure and quite hard to get vocaloid english to sound good or Pronounce/Accent right. Which is completely opposite to the Synth V English bank that are much more mature and more friendly to edit. You might not believe what i say but this alone contributes such a huge impact of how english voicebank will "end-up" sound like. Just simply easy to edit can make a Night/Day difference for a singing performance.
Singing range and tempo range is just a "suggestion" and should be treated as such, most VBs can be push far beyone the limit before it sound like a noise.
Honestly anything will sound like a beeping sawtooth wave if push unreasonably excessive.

There is always some unwanted / leftover weird noise when exports. Some VBs are doing excellent, some are doing poorly. So what came out from any singing synth engine are not absolute. This is where mixing stage are for.

 

[sunnyp4rk]

I'm convinced the word "vocalsynth" was invented by the Big English Speaking Community as a way to continue selling Teto Is Not A Vocaloid to new fans.

Or maybe people use it because there are multiple vocal synths now? Are you also bothered by the fact that Teto is Literally Not a Vocaloid? She was never released on Vocaloid, and a concept voicebank does not count.

There's no point in using Vocaloid to describe every vsynth when it's not the only one anymore. It would be weird to call SynthV characters Vocaloids unless they're actually on Vocaloid. I think that separation is a good thing actually.

Vsynth is the perfect term because it describes what these pieces of software are, but it doesn't describe them with a specific brand name. I'm not going to call UTAUs Vocaloids unless they actually are on Vocaloid, like with GekiKaze V6 and Ooka Miko and her upcoming Vocaloid6 voicebank. I could call them Vocaloids because they're actually on there.

I would think vsynth is an easier term for new fans to understand, rather than calling everything Vocaloid.

 

[MagicalMiku]

It's an interesting fact for both the fan community and the companies: Crypton, for example, on their blog and twitter often refers to them as Piapro Characters or virtual singers, but when is another character from, let's say Internet Co, they write something like "Gumi from Internet Co.", and if necessary, adds more details about the voicebank engine used for the song or the artwork.
Virtual singers is also the "official term" used for Project Sekai, and it's interesting that on Sonicwire official shop, all Piapro Characters (both Vocaloid and NT) are listed as "virtual instruments".
How about on Yamaha Vocaloid website? Well, there is another term, voice bank:

Voice Banks are libraries used in the VOCALOID Editor to synthesize singing voices. Choose your favorite character from over 100 different characters.

 

[Grzesiek11]

[...]

tfw you post an actually unpopular opinion in the unpopular opinions thread

Are you also bothered by the fact that Teto is Literally Not a Vocaloid?

No, I'm bothered by the fact apparently everyone in the English speaking community loves to correct people when they call [insert voice not on the VOCALOID™ engine] a "vocaloid".

There's no point in using Vocaloid to describe every vsynth when it's not the only one anymore.

Why? It's a better sounding word than "vocalsynth". Also vocaloid* not Vocaloid.

It would be weird to call SynthV characters Vocaloids unless they're actually on Vocaloid.

Why? It's perfectly normal for trademarks to become genericized. List of generic and genericized trademarks - Wikipedia

I think that separation is a good thing actually.

Why? Engines only matter when discussing engines. Most of the time, you are not discussing engines.

Virtual singers is also the "official term" used for Project Sekai [...]

Oh wow I forgot about that, this might actually be worse than "vocalsynth" because it's also independently used to describe what you could describe as "singing VTubers" (obviously there are normal VTubers that sing, but this is kinda a separate category, from what I understand), which is a pretty much unrelated phenomenon. バーチャルシンガー - Wikipedia

 

[sunnyp4rk]

Why? It's a better sounding word than "vocalsynth". Also vocaloid* not Vocaloid.

Well I'll be petty now and use the proper official spelling of VOCALOID now because you want to correct that. Also, it's not a matter of "what word sounds the prettiest" its "what word can convey what a vocal synthesizer is the best". yes VOCALOID was the standard for ages but it's not anymore. Vocal synth literally means vocal synthesizer. VOCALOID doesn't actually mean that. It's just the ENGINE name.

Why? Engines only matter when discussing engines. Most of the time, you are not discussing engines.

I mean, I specifically do discuss engines. Macne Nana is a VOCALOID and UTAU because she was on both engines. If she wasn't on VOCALOID, I wouldn't call her a VOCALOID.

I'm not going to be like "oh hell yeah my favorite VOCALOID chis-a!" because she was never ON VOCALOID. It doesn't make sense. Like I think lumping every single character into the term VOCALOID would make it more confusing for me if I was a new fan, because not every character is on VOCALOID.

edit: I'm sure there are probably people in the Japanese speaking sphere of the fandom who are also just as picky about stuff like this. It's just not something you'd see unless you were in that sphere. I don't think its only Big English Fandom that would be super focused on correct terminology if you looked around.

 

[lIlI]

Reminder to keep it civil when discussing singing synthesiser semantics, everyone.

 

[MagicalMiku]

I think most of the confusion of the users come from how and where they learn about the characters.
Let's say that most of users know about Miku, Teto, or any other popular character, from Project Diva (there is Teto's extra DLC), from official concerts, or from popular songs available online.
If a user knows from Project Diva or Project Sekai, probably will think that Sega is actually behind all of those characters. Which, to be honest, is someway true: Sega holds all the rights of Miku games on all platforms, developed by Sega and other companies. Sega also is responsible for making the concerts, with its company Marza Animation Planet. And Sega designed many of the original modules for both Diva and Sekai (others were based on official illustrations/song or Piapro art contest).
Now, I feel there are two kinds of user: user n.1 likes those songs and characters, but not so much by investigating more about them. They are Vocaloids made by Sega. Yes, that's it. But for user n.2 things are different: will research more about them, on official websites, wiki, and so on, and will understand the various differences, the different engines made by different companies, why Crypton has both Vocaloid and NT, why Teto, IA and Gumi have different outfits for different engines. And will learn about new characters. So it's ok and a good thing to correct new fans, especially if it's user n.2 type.
Vsynth sounds very good (and technical), vocaloid of course is more iconic. With so many different engines available now, I usually just say [character]+[engine], is just easier for me. For example, Miku V6 or Miku NT2, or if I want to compare some of her previous VB (voicebanks), I say Miku NT (which can still be used, if you have the installer), Miku SP (Super Pack), Miku V3, Append, ..
I think Yamaha, Crypton and the other companies should translate some of the books that are available in Japan and that can explain very well the history of voice banks.

 

[IO+]

I honestly don't really care what people gonna call it he, she, it or vsynth, vocal synth, VBs voicebanks, vocaloid, utau, cevio, synthv, or whatever. Because these are sub category.

I actually like when people call it vsynths / vocalsynths because it generalized and wrap them up together in the low processing fashion and that's what i call efficiency. But if people wanted to be more specific like talking about vocaloid engine, obviously they can use that sub category. I see no point to take things like a brickwall limiter. From an information-processing perspective, using "vocal synth" as the umbrella term is efficient because it avoids unnecessary classification until classification becomes useful. That's basically how taxonomy works in most fields.

Honestly.... Some VBs have multiple version across engine. Zundamon for example.. She's already exist in these engine likeUTAU, VOICEVOX, NEUTRINO, VOICEPEAK, Seiren Voice, CeVIO AI, VoiSona and god knows many more in the future...

So when people talking about Zundamon but which zundamon? As you can see..... this is where sub catagory are very useful. Where discussions often become unproductive is when people treat the subcategory as if it's the primary identity.
Example: "Teto isn't a vocal synth, she's an UTAU." But UTAU is the software platform. Teto is still part of the broader vocal synthesis ecosystem.
So Use "vocal synths" term when discussing the ecosystem as a whole.

So basically.
- Use "vocal synth" when discussing the ecosystem as a whole.
- Use the specific engine when the technical distinction matters.
In my opinion it's fairly sensible way to communicate because it minimizes unnecessary detail while still allowing precision when needed. The existence of characters like Zundamon across UTAU, VOICEVOX, NEUTRINO, VOICEPEAK, CeVIO AI, VoiSona, and future engines is actually a good example of why both levels of classification are useful rather than mutually exclusive.

This is my opinion though, agree and disagree are also opened for discussion.
I believe that neither position is inherently wrong. They're optimizing for different things like simplicity versus precision. The real question is whether the distinction changes the meaning of the conversation. If it doesn't, the broader term usually works fine. If it does, the subcategory becomes valuable.

 

[Grzesiek11]

Well I'll be petty now and use the proper official spelling of VOCALOID now because you want to correct that.

The point is that using the official (or uppercase-first) spelling is wrong, as the word I think is best to use for the broader category of vocal synthesizers is "vocaloid". Specifically lowercase, because it's not the name of any specific engine.

VOCALOID™ - a specific vocal synthesizer engine made by Yamaha
vocaloid - some vocal synthesizer

Also, it's not a matter of "what word sounds the prettiest" its "what word can convey what a vocal synthesizer is the best".

Words convey meaning by existing and being used, not by what specific letters or other words they are made of.

I mean, I specifically do discuss engines. Macne Nana is a VOCALOID and UTAU because she was on both engines. If she wasn't on VOCALOID, I wouldn't call her a VOCALOID.

When discussing engines, in which way you use the word "vocaloid"/"VOCALOID™" is clear from the context.

"The recent vocaloid song by XXX featuring IA" (even if it's the CeVIO voicebank)
"The recent vocaloid song by XXX featuring IA V3" (it's definitely not the CeVIO voicebank now)
"IA recently got a new VOCALOID™ voicebank" (clearly you mean the VOCALOID™ engine, even if you are speaking out loud or write everything in lowercase)
"She's a vocaloid producer" (maybe using SynthV)
"She produces songs using VOCALOID™" (now definitely not using SynthV)

What do you think of the words "photoshop" and "google"? Is it weird and confusing if I've "photoshopped" an image with GIMP, or "googled" something outside Google? Would the word "immanipped" be better, because it combines the parts of "image" and "manipulate"?

The phrases "vocal synthesizer" and "image manipulation" are normal phrases to be used in professional contexts, but that doesn't mean "vocaloid" and "photoshop" are worse than "vocalsynth" and "immanip" in a casual context.

That said, I agree the distinction between the broad "vocalsynth"/"vocaloid" category and the subcategories of specific engines is often useful, and I'm not trying to argue that it isn't. My point was that I don't think it's important that the name of the broader category must be completely separate, though I guess I didn't get it across very well.

 

[wrong_thyme]

my hot take is that vocal synth just sounds better and i'd sooner chew barbed wire than call non-vocaloid vocsynths vocaloids

 

[sunnyp4rk]

my hot take is that vocal synth just sounds better and i'd sooner chew barbed wire than call non-vocaloid vocsynths vocaloids

Same lmao. vocal synth is clear and concise and covers all of them rather than just one engine (like if I'm talking about both UTAU and Synth V, it makes more sense for me to say vsynth rather than VOCALOID, right?)

Words convey meaning by existing and being used, not by what specific letters or other words they are made of.

Is that not literally what I was saying... Vocal Synthesizer (or vocal synth) is the meaning here. It just describes what these programs are. VOCALOID doesn't actually describe the meaning of a program. When I hear VOCALOID, I think of specifically VOCALOID. I don't think of UTAU when I hear it, because they are literally two different engines with two different names.

Would you loop something like Voiceroid into the term VOCALOID as well? Talk synths (Voiceroid, VoiSona Talk, AI Voice) don't operate the same way as VOCALOID. Vocal Synth covers both song synths and talk synths, because its just a simplified blanket term to cover everything. You don't HAVE to specify whatever you're talking about. I think it's literally the easiest way to talk about vsynths with new fans, or even old fans! It's just simple.

I don't know why we're opposed to simplifying things for new fans???? Like I think if I was a new fan in this modern day and age where there are many more
popular synths than just VOCALOID and UTAU (CeVIO existed in 2018 but I barely heard anything about it at that time), I'd be really, really confused if everyone kept saying that every character was a VOCALOID (because they aren't).

Like IO+ said:

Use vsynth when talking about everything, use engine name when talking about the specific engine a character is on. Like I will use Kazehiki UTAU or Kazehiki V6 depending on which version I'm talking about. But I'm just going to say "my favorite vsynth character Kazehiki" because engines don't matter in regular speech about the character.

"The recent vocaloid song by XXX featuring IA" (even if it's the CeVIO voicebank)
"The recent vocaloid song by XXX featuring IA V3" (it's definitely not the CeVIO voicebank now)
"IA recently got a new VOCALOID™ voicebank" (clearly you mean the VOCALOID™ engine, even if you are speaking out loud or write everything in lowercase)
"She's a vocaloid producer" (maybe using SynthV)
"She produces songs using VOCALOID™" (now definitely not using SynthV)

This would confuse me as a new fan. If someone is using Synth V, they shouldn't be using the term VOCALOID, since it isn't a song using VOCALOID. Only use the term VOCALOID when using the actual engine VOCALOID. And if I'm talking about a character getting a new voicebank on a specific engine, then I'd use that engine's name anyways! "Recent VOCALOID song by XXX featuring possible voicebank on another engine" is also wrong because CeVIO isn't VOCALOID. It's CeVIO. It's own thing. It should be labeled properly with it's own engine name. This will just confuse people. I use both GUMI VOCALOID and Synth V and I specifiy in my cover titles which engine I'm using.

We need to let go of using VOCALOID as a blanket term. We have other synths now, it's done. Over. VOCALOID isn't the only one now. We have at least 6+ other engines and continuing to use VOCALOID as the blanket term is just dumb.

(This is the last I am replying to any of this. I am too autistic and passionate about dumb things like categorization and could argue for ages, so I am cutting myself off)

 

[IO+]

All voicebanks are just part of the vocal synthesis ecosystem. Fact won't change.
So Use "vocal synths" term when discussing the ecosystem as a whole is make senses.

When in discussion i much prefer efficiency and minimizes unnecessary detail while still allowing precision when needed.
- Use "vocal synth/Voicebanks" when discussing the ecosystem as a whole.
- Use the specific engine when the technical distinction matters.

 

[bibbs]

voca-semantics is serious business.

here's an unpopular opinion: i think the "vocal synths aren't generative ai" is a losing argument. i don't think the point should be "generative ai bad, other ai good," it should be about *consent* and *ethics*. an ai voice bank is basically training data for a particular voice, sourced from a consenting voice provider. voice banks have been cancelled when the voice provider got uncomfortable after community reception, or the engine sounding too realistic before. in utauland (technically not ai) they have terms of use, written up by usually the voice provider themselves, enforcing rules around violent content, sexual, impersonation, commercial content, etc.

on the other hand, the suno and udio's of the world are sourced by scraping any and all data from the internet without consent, stealing most often from small, independent artists and anyone's voice it happens to come across.

third, vocal synth engines exist as musical instruments, without any understanding of how to write music, you can't use them effectively. there's no "make a rock song with lyrics" type of prompting involved. the ai exists as a tool to enhance the ability humans already have, rather than to replace humans entirely.