New languages I would love to see added to SynthV

SynthV is already compatible with English, Japanese, and Mandarin. If there’s any new languages I would love to see added, these are what I suggest:

  1. Spanish/Español: One of the most popular languages besides English. VOCALOID is already compatible with Spanish, but VOCALOID voicebanks directly meant for Spanish are pretty outdated. (no new voicebanks have been made for Spanish since VOCALOID3)
  2. Korean: Again, VOCALOID is already compatible with this language, but theres very few voicebanks currently designed for it.
  3. German/Deutsch: It’s perhaps the most popular Germanic language besides English.
  4. Russian: It’s the most popular Slavic language, and i’ve seen some vocal synth music made with this language.
  5. More Romance languages (like Portuguese, French, and Italian): They’re also quite popular languages. (Although I wonder how the program would handle the nasal vowels of French and Portuguese)
    I’m very curious about how SynthV’s AI voicebanks would handle the addition of such new languages, if they ever get added. Are there any other languages you suggest should get added to SynthV in the future?
11 Likes

This post was flagged by the community and is temporarily hidden.

1 Like

The Spanish Vocaloid community is kinda known for its toxicity but as long as people can get around that I think that’ll help SynthV a lot. Likewise, the Korean Vocaloid scene is kinda infamous of its own fair share of controversies so it’s something to tread carefully of but if it happens it could really pay off. I do vouch for French though.

(I’m not native English speaker, sorry for potential mistakes. )
As for the Korean language, I think it is possible to mimic the pronounciation of Korean by a finetuned SynthV dict (Korean → Chinese Madarin), because visemes of Korean is a subset of Chinese’s. Especially for SV Studio after ver1.5.0 which brings cross-lingual support, AI banks in Chinese mode seems to have more phonemes than previous standard banks (for example, not only a but also A in X-SAMPA). Though some strange pronounciation may take place, they can be manually corrected.

LOL is that so??? I wouldn’t really attribute any of those controversies to particular linguistic ‘sides’ of the vocaloid community… I mean hasn’t the English and Japanese side of the fandom seen it’s fair share of fandom drama / controversies?
Imho, I think that when a fandom starts to look like it has lots of toxicity or w/e, that it’s really just a symptom that comes with having a highly active online community.

I propose the French. I live in Quebec, Canada, there are more than 50% people who speak French around me.

1 Like

I think you meant to write phonemes. The viseme is the shape the mouth makes while forming the sound, while the phoneme is the sound.

But good point! :+1:

I also like french :slight_smile: spanish would be cool too but i would probably not use that??

French !!! I’m French from France and my voice is terrible, I hope to hear Kevin/Kyo/Muxin in French voice to sing for me :slight_smile:

This post was flagged by the community and is temporarily hidden.

This post was flagged by the community and is temporarily hidden.

I’d throw my two cents in for German language. German lyrics are quite hard to achieve by English/Japanese/Chinese phonemes. Some might he covered by French phonemes, though, so this might be a big step towards it, too! :grinning:

A vote for operatic Italian voices, that would be great…

This post was flagged by the community and is temporarily hidden.

This post was flagged by the community and is temporarily hidden.

1 Like

Japanese side I’m not sure, but English, yes, but it’s nowhere near as bad as the Spanish/Korean side of things. I wasn’t going to mention any of them because it’ll cause flame wars, but let’s just say one guy wanted to make a set of Spanish voice libraries but his compatriots have been criticising his libraries for sounding too similar to some of the existing libraries, which is kind of inevitable with how much development the engines have yet to undergo (you can technically make any two libraries sound like each other if you put in the right amount of work and research). And then the guy (understandably) lashed out at the rest of the community and announced his resignation. As for the Korean side, way worse. There’s been a voice provider that’s been known for some… yea, I can’t say what she did without risking getting banned from all of the internet, we all know there’s a good reason why all platforms have strict policies against NSFW content.

English side of things on the other hand, the worst it’s ever gotten was like, what, voice providers of a couple Vocaloids treating their characters like darlings and kinda being aloof from the rest of the fanbase (to be personally fair I’m on the middle of the fence about it)? Or do you want to talk about how I wound up banned from several platforms over defending several libraries’ viability in rock and metal, because that’s the biggest controversy I can think of other than the aforementioned darlings controversy… (because surprise, I’m a Voca-P that’s been around for at least six years in the scene!)

Also, toxicity is exactly why more and more people have been leaving forums and servers, not even counting the number of people that wound up wrongful-banned from several platforms, myself included.

Not trying to say Spanish and Korean should be avoided, but there’s been some massive scandals that took place in those communities and that’s why the fandom there are going to be really sceptical. That said, we may never know, because there have been some massive controversies in EN too and half of them happened only because I was there.

Again, like I said earlier, French is one of the languages I really do want to push for. I’m also thinking Estonian.

Another thing, like, if anything, I for one, like, would still vouch for Korean despite the R-rated scandal that happened not long ago, because I would definitely love to try and be that bilingual connector between EN and KR (Yes, I’m bilingual), and I’ve seen what the people have been doing in that scene and I would definitely love to try and get to know more of the scene.

I suspect that there are three main factors involved in additional languages:

  1. The market demand for the language. If there’s enough profit to be made, it’ll be made. But “enough profit” is a relative thing. If a Spanish voicebank brings a return of 5x, but a Japanese voicebank brings in a 10x return, you know where the smart money is going. (Of course, this ignores the ability to grow a market).

  2. The cost for Dreamtonics to make the voicebank. If an outside investor is willing to take on the cost of making a voicebank, the cost and risk for Dreamtonics drops greatly (there’s always opportunity cost of putting resources elsewhere).

  3. Voice data for trans-lingual synthesis. Once an initial voice for a language is available, the raw material for training that trans-lingual synthesis exists.

So I’m guessing that the most likely way for a new language will be added to SynthesizerV is for an investor to partner with Dreamtonics. Once that’s done, it’s likely that trans-lingual syntheses could be done using that voice data and existing voices to teach SynthesizerV how to perform trans-lingual synthesis to that target language.

In theory, and existing voicebank in some other language could be used as source material.That is, existing voicebanks could be used to produced training data.

On the other hand, the output wouldn’t be better than concatenative synthesis, and I suspect that to produce the highest quality results, Dreamtonics would want to record the voices from scratch.

After all, if you’re going to invest all that money, you’ll want the highest quality product you can get, right?

1 Like

Hm, I doubt a German-implementation would be successful - unless they’d made commercials everywhere to let us people know about SynthV, to be precise.

(You think German-me saw at least some Hatsune-Miku-stuff here in local stores? Unfortunately: Nope, never, as long as you don’t count a single rhythm-game for NDS3. Figurines? CDs? None. :sighs:)

As for imitating the language with our current options…
Though I used no dictionary for SynthV’s pronunciation - which means in other words, that I merely wrote “gibberish” text - our Chinese girl Minus (when compared to a veeeery old Vocaloid project of mine) did not the worst job in singing “Warten, dass ein Wunder kommt” from Encanto’s German version ( YouTube link ).