my opinion on miku for synthv

So i have seen alot of people say that they think that miku will get a synthv voicebank and heres what i think : i think that crypton will keep trying to make updates for NT but lets be honest its really about the money, so once they see that synthv is getting more sales, a better image, and is more popular they will put her on synthv, but thats just my opinion

1 Like

I can’t imagine it. Miku can’t get a realistic AI voice because it won’t be Miku anymore. I rather think that they will improve their engine to restore the sound of her voice to that known from vocaloid voicebanks, only in better quality.

Dreamtonics will somehow “improve” their rendering engine by making less realistic and more like a Vocaloid? :crazy_face:

People who want the Vocaloid sound of Miku already have it. That market’s already been filled.

The point of porting a voice SynthesizerV is to take advantage of what SynthesizerV offers - a more realistic voice with a more human performance.

Of course, Dreamtonics will do whatever they choose, without listening to me. Heck, they could prove me completely wrong and you completely right. Time will tell. :slight_smile:

Dreamtonics will somehow “improve” their rendering engine by making less realistic and more like a Vocaloid? :crazy_face:

Nothing needs to be “improved”. Synthv still has a classic mode, which we almost forgot about since all recently released voicebanks use AI technology.

People who want the Vocaloid sound of Miku already have it. That market’s already been filled.

The problem with Miku is that Crypton abandoned the VOCALOID engine in favor of its own solution. Although there were many good arguments for this decision, technologically the NT engine is outdated and sounds even worse than Miku V2. It is also unlikely that Crypton will change its policy and, after resigning from Yamaha’s services, enter into cooperation with any other company. Even if the sales of voicebanks dropped due to poor quality, they probably don’t care, because the real money they earn from Miku comes from other sources, such as gadgets, games and gigs.

The question is how long will it be possible to buy Miku V4x and how long will old voicebanks be compatible with the available Yamaha software? Probably VOCALOID 7, if released at all, will drop support for the old technology (as they already did with V2) and will be something completely different, based rather on what we now know as VOCALOID β-STUDIO (which looks great by the way).

Miku V4X comes with Piapro studio, so compatibility with Vocaloid is a non-issue. V5/6 are already worse than V4 was when it comes to using Miku, due to the lack of cross-synth. Of course Crypton could discontinue V4X, but they’d be silly to do so while NT is still poorly adopted.

I don’t think there’s much point discussing the Japanese capabilities of Miku since V4X does what almost every Japanese user needs it to, minus having a rather unfriendly editor to work in (at least by SynthV’s standards). There’s just not much room to improve when the V4X sound is what people want.

It’s easy to say “just make a SynthV Standard voice database”, but I’m not convinced that would help anything aside from providing a better editor to work in. The only English Standard voice released was Tsurumaki Maki English, and if you’ve ever tried to make her pronounciation remotely as good as her AI counterpart… well, it’s not much easier than using Miku V4 English, at least from my experience.

I’m not counting Eleanor Forte lite here because her voice provider speaks English natively, so it’s not a helpful comparison. I think English users may have gotten warped expectations about Standard voices based on how easy Eleanor Forte lite was to use compared to English Vocaloid options, but remember that AHS stopped making the Standard versions of their voices because very few people actually used them; it’s likely that SynthV’s concatenative capabilities aren’t as impressive as we might think if we’re applying our expectations from Eleanor to a Standard voice database with a Japanese voice provider.

I admit that for a long time I wasn’t convinced AI voices could capture vocal tones as well as multiple style-specific Standard voices could (ie. Crypton’s V4X packages or GUMI V4 Complete), but I feel I’ve been happily proven wrong as the software has continued to mature.

I think the important thing to acknowledge is that each additional vocal color for a concatenative voice requires recording and developing an entirely separate product, which increases the cost dramatically. If the same amount of effort and budget were dedicated to each of an AI voice’s vocal modes, I no longer doubt they would be able to achieve the same level of variation. The main difference being that with AI vocal modes it’s not all-or-nothing; you can still get close without having to literally double, triple, etc. your development costs. Sure, Solaria’s Power vocal mode is probably not as powerful as if they had released a dedicated “Solaria Power” voice database, but they were able to achieve it with a budget that wouldn’t have allowed for two separate products.

But, back to Miku. Is there any point making a SynthV Standard voice? Probably not. She wouldn’t sound the same, and might be slightly easier to use, but in order to match V4X they’d need to develop and release multiple voice databases, which would be extremely expensive. If they took the same approach as NT and only released a small selection of vocal colors, we might not get an English voice database at all.

What about an AI voice database? Well, I’m sure it’s possible if Dreamtonics made it their express goal to figure out how to capture Miku’s sound as an AI voice, but that would take dedicated R&D effort that would diverge from their existing priorities. As-is, it’s been discussed to death in many places, but just making an AI voice with the same voice provider would likely not meet anyone’s hopes or expectations. I’m sure many anglophone users would prefer it thanks to cross-lingual synthesis, but unless they really nailed the vocal timbre, the Japanese market would probably still stick with V4X.

All that said, if they did release Miku for SynthV, I’d be one of the ones to buy and use it, at least in conjunction with V4 English. I just don’t think it has the potential to be “the next big thing”. As usual with these sort of speculations though, I’d be happy to be proven wrong.

crypton is selling synth v now so it may mean partnmer ship