Swedish Singer YOHIO to Voice New English-Japanese Sound Bank

On his official Twitter account, Swedish singer YOHIO has confirmed that he will be voicing an upcoming English-Japanese bilingual voice bank for music software company PowerFX, the company behind English VOCALOID voice banks Sweet Ann, Big Al and Oliver. PowerFX CEO Bil Bryant also confirmed with us the news, stating that recording had already begun, although pricing and other details have not been set yet. The effeminate-looking YOHIO, currently 17, has been featured on Japanese music news site BARKS and was influenced by visual kei culture. He had already released a mini album in Japan last April and is set to release a full album this April 24 through Universal Japan. Rumors about PowerFX making a bilingual voice bank with a Swedish singer (as well as subsequent plans to make an American sound bank) were sparked by an email from Bil Bryant to a customer that was posted online before the parties involved were ready to properly announce the project. The posted email also stated that PowerFX was two months away from a release.


30 thoughts on “Swedish Singer YOHIO to Voice New English-Japanese Sound Bank”

  1. Wow, new Vocaloid, and a new male V3 in japanase! yeah! finally!
    I like the voice of this singer, that could be my next favorite virtual singer!

  2. That’s the first smart move I have seen PowerFx make in awhile. Maybe They will start to change my mind

    >as well as subsequent plans to make an American sound bank

    Never mind.

    1. How is making an “american sound bank” bad? It would be a nice relief from all the british loids ZeroG has released so far. Plus I think the only american accented vocaloid there is so far is Big Al and he sounds pretty good…

      1. ….Are you serious? American accents are retarded when it
        comes to recording. (And generally speaking especially the south) Also, X-Sampa
        wasn’t built to do American English to start, it had to be modified for America
        and it still sounds like mud because they always take short cuts and mix King’s
        English with American English which you do not do under any circumstance.
        Seriously “earn” and “urn” are not the same in
        pronunciation and if you think they are, you seriously need to go relearn English.
        (I know I have used this example before but it fits so well in the short cut
        problem.)

        1. I’m confused by the example for ‘earn’ and ‘urn’. These are homophones, at least in American English. Various Google searches for pronunciations list them both as being identical.

          1. As someone who grew up in the US, I wasn’t aware that they were supposed to be pronounced differently, either. The Oxford English Dictionary apparently lists both words as: /ɜːn/ and Merriam-Webster lists both words as: ˈərn I don’t think I know of any sources more authoritative than those two with regards to British and American pronunciation…

          2. Both are somewhat correct but it depends on the sentence itself. /ɜːn/ (urn) is used when you are talking about a vase that holds ashes or water. ˈərn (earn) is when you are entitled to something that you have worked for. Also the Oxford site would probably be more accurate since it shows both enunciation of British and American english.

          3. I’m using the Oxford Advanced Learner’s Dictionary. The one that yours is linking to the Oxford Advanced American Dictionary. You need to make sure you have the oald8 in front of it If you google it is the first site that pops up.

          4. The one I’m looking at the one with the link above says they are different and I can clearly hear the difference.

          5. Would you care to reread my comment and your response? What you’ve written indicates that you are conflating the difference between the pronunciation of “earn” and “urn” with the difference between the pronunciation of these words in American vs. British English.

          6. Actually ərn appears to be rhotacized whereas ɜːn is not. American being a rhotic accent, pronounces these both with rhotic vowels (both as /ɝn/) And the majority of British accents pronounces them with non-rhotic vowels. (As either /ɐːn/ or /ɜːn/)

            Nevertheless, the accent in the long run does not matter much when recording a voicebank. The pronunciation is not affected by the notation in any way even though they may not accurately match. Vocals produced will sound natural except for certain cases where the accent recorded posseses multiple allophones for a vowel when RP does not. (For example the American pronunciation of æ being pronounced as eə before nasal consonants.) However this seems to be the only problem, (other than the lack of an alveolar flap in the program for American pronunciation) that accent differences has caused in English vocaloids.

        2. Southern accents are fine and there’s literally nothing at all wrong with their phonetics. As the sound V will have a different infliction than a British accent they’re both pronounced “uh”. Also please do not use ableist words such as retarded. They’re extremely offensive. Thank you.

          1. You think there is nothing wrong with slurring words and
            using inappropriate tone inflections? Why do you think so many comedians overseas make fun of North American accents and even American’s make fun of southern accents? It’s because it is extremely lazy and usually sounds like someone who has been bashed on the head one too many times.

          2. “Bashed on the head one too many times”. You’re really really, really ableist you know. And anti-anglo, but that’s besides the point.

          3. Oh well that what the American and Southern American accent tends to sound like most of the world. Get over it.

  3. Ahaha, my sister is going to freak. She absolutely loves YOHIO. //unimportant

    Well, since YOHIO has some experience in singing in other languages, maybe, hopefully, this voicebank will have better phenomes…? (I can hope, can’t I?)

      1. Sorry, can you please clarify? I assume you mean “bad” shortcuts, correct?
        I’m just hoping that faint hope that the new Engloid will become “better”. I don’t know how it will become “better”, since I’m not exactly experienced in using Vocaloid (or phenomes, even). But if they eliminate using “bad” shortcuts, I would consider that as “better”. If you don’t mind, can you explain how eliminating “bad” shortcuts would be “unfortunate”? (Or correct me, if I completely interpreted your comment wrong…)

        1. I mean PowerFx and Zero-G take many bad short cuts when they make english vocaloids. (But for some reason all the companies that make english banks take bad short cuts)

          Though you mentioned you don’t quite understand phonemes just think of them as sounds represented as symbols in the engine.
          To correctly program a basic english voice bank in Sampa or X-Sampa (The coding Vocaloid engine understands) you need the following vowels . @, a, A,Q, E, V, O, o, @, 3′, e, I, U, u, i, eI, aI, OI, @U, aU, I@, e@, and U@. (If you want an Australian accent you also need to add } into the list)

          However this is the vowel chart that is in the english banks @, V, e, I, i:, {, O:, Q, U, u:, @r, eI, aI, OI, @U, aU, I@, e@, U@, O@, and Q@.

          As you can see (Or at least count up) there is a few missing which throws off the banks completely.
          PowerFx and Zero-G try to make up for this by making certain vowels pull double duty to make up for the vowels they eliminated.
          The problem with this is that which makes it so that the engloids have a hard time speaking even basic english at times.

          Also there is the problem of trying to have other phonemes like @r replace 3′ which the two sound nothing a like so words don’t sound correct when placed in. Also there is the problem of the incorrect dictionary that tries to replace @ with V when that is not correct at all. Finally there is the problem that they did not program the banks to detect homophones which lead to more incorrect pronunciation.

          1. Ah, okay, so the problem is that English voicebanks are missing all those phenomes and replacing them with others, as well as that problem with homophones.

            I can see that you’re very knowledgeable in this topic, thanks for taking the time to explain that to me!

            So, these problems (the missing phenomes and undetectable homophones) are some of the bad shortcuts that PowerFX and ZeroG practices, right? So if they start including these missing phenomes and address the homophones issue (and I’m not saying that they will), would that be considered as a step in making English voicebanks “better”?

          2. Natasha Celestine

            Yes if the companies did start adding in the missing phonemes and fixing the homophones issue it would help the english vocaloids a lot. However I don’t ever see it happening.

            As for me being knowledgeable I mostly just dig through archives of sites which anyone can do. (which why people don’t is beyond me) However, it has lead to me not being liked at all (Which I don’t mind. I prefer playing devils advocate) because I am very hyper critical of both the western companies and the western fandom overall.

            I like applying the old saying of “If you are gonna do something at all you might as well do it right.” Which the west still clearly lacks when it comes to vocaloid.

          3. Actually I’m pretty sure the vocaloid producing companies have no say so in the phonetic mapping of languages in the program. These systems are all developed in Spain (Where the program was also originally developed) and then approved by Yamaha in Japan, which is why the phoneme mapping for the other three languages in Vocaloid, English, Korean, and Chinese, is so difficult to use and really limits or complicates the way these voicebanks are produced.

          4. @Syo

            Well yes and no. Originally SAMPA (Followed by X-SAMPA to Improve on the build) codding language was developed in 1989 as a code work around for IPA so that a user could simple type in a language for a computer to read. Pompeu Fabra University picked it up after words because it seemed to fit what Yamaha was asking for when creating vocaloid. (At least that is the jests of what I get from there thesis on vocaloid) So companies do have the ability to create the required sounds because they have the entire dictionary at their finger tips (and anyone else at for that matter) by just going the SAMPA site and using the software they provided free of charge. The main problem that is Yamaha decided to try to take short cuts on the programing (and even some of the companies in sound recording) to cut down cost which leads to the english banks being so terrible in quality. However, when ever it is brought up to the english companies asking them to tell yamaha that the short cuts are ruining the product they try to side step it while the japanese companies actually take the complaints head on. (Like how with Sonika when Zero-G just said that they purposley made her a low quality bank when they advertised her as a high quality pop singer to start or how SBS claimed SeeU was a trilingual vocaloid)

Leave a Reply

Your email address will not be published. Required fields are marked *