SynthV the great dreamtonics christmas announcement theory zone

[BANKARA_TEEvie]

So, we've had a good few big updates to SV revealed/teased/released on or around Christmas a few times now - NineZero, the XLS update, even SV2 getting teased.
I have reason to believe this will happen again this year, because while SV2 has only just released, I feel like we don't really know the kind of direction that the engine will be going in (that's to say, we can assume they'll keep supporting it with major feature updates like sv1, but what can they really change!) - what sort of changes we can expect in the very long run, as the SVStudio that existed on launch is very different from how it stands today after SV2's release.
could we have another complete engine-overhaul like hdvm? could we get even *more* languages? or perhaps a feature we don't even know could be supported/added, like realistic growl effects for EVERY voicebank. NONE of these feel very likely, which is why I'm asking the question about what people think could be revealed! Or, at the very least, what ways DT could feasibly do to keep updating and upgrading SV2.

Me personally, I think an update that focuses on adding more depth to XLS would go a long way... more language support sure, but I think now is time more than ever for some kind of way to finetune a voicebank's accent when singing in a different language.

is this a boring discussion? a BAD thread? who care! let's talk about it :)

 

[lIlI]

I find this topic really interesting to think about, it definitely gets exciting towards Christmas to see what crazy thing Dreamtonics drops. I'm tempering my expectations a little this year, as SV2 is still recent enough that their schedule may be too busy for a huge Christmas announcement. My conservative guess would be the full release of version 2.2, which is currently in beta - but that wouldn't be that exciting as we've already seen most of what it'll have to offer. New letter voicebanks are regular drops, so I'm not sure about that either... My far out guess would be French, which has been floating around as a possibility for a while.

What I'd really like is just some blanket improvements to areas where SV2 voicebanks fall behind SV1, I still haven't switched over to SV2 despite owning it because 90% of my voicebanks are reported to be weaker/buggier on the new engine. But I wouldn't be opposed to a big surprise! Maybe the addition of a famous character?

 

[BiscuitsandTriscuits]

In a perfect world of my own creation they will just drop the bomb of Crypton expanding the engines their vbs are one with Luka and Meiko to start (and also for some reason a sweet ann update from powerfx's grave) but I'm expecting maybe a real release for the 2.2 and some letter vbs and maybe a new ahs character maybe

 

[Bookworm2]

Methinks... I dunno. There are TONS of things I'd love to see from Dreamtonics, but most of them are wistful thinking. Like I'd love for them to add XLS for Latin, but I just don't think that'll ever happen. Realistically, I feel like Growl/Vocal Fry as a feature for all voicebanks might happen, but I'm not too sure about that. Likewise, Glottal Stops being added would be really nice, but I'm simply not sure about that. I guess maybe an update to the Rap mode to make it more like actual rap and less like slurred singing could be possible, as I'm pretty sure that's a common point of complaint among users looking for a true Rap vocal synth. On that note, maybe a talking mode? I'm not sure how highly requested that is, but I guess it could be possible.

Keep in mind, most of this is either wistful thinking or theory making, so take it with a grain of salt. However, it would be nice if Dreamtonics did any of this wink wink nudge nudge.

 

[mary34]

Cowboy SynthV voiced by P.M. Seymour trust

In all seriousness it's probably gonna be something relatively modest, I imagine they were busy this year with many things. Maybe something crazy happens and we get French but like? Idk tbh

 

[Rylitah]

On that note, maybe a talking mode? I'm not sure how highly requested that is, but I guess it could be possible.

Probably not since Voicepeak exists. My dream is for Voicepeak to have XLS of its own though. I have wanted for years for talk synths to have more English support because people already like making talkloids! JP users for years have been making entire dramas with Voiceroid and other talk synths, and I really think creatives in the Western fandom would thrive too if they just made using these TTS programs more accessible.

All I want for Christmas is English Voicepeak make it happen Dreamtonics

I don't have opinions or wishes for Synth V besides Minase Kou SV so this is all I can contribute to this thread lmao

 

[parallax_fifths]

Dreamtonics will announce a $10bn investment from nVidia to improve their technology, which will be used to rent $10bn of computational resources from OpenAI, who in turn will buy $10bn of hardware from nVidia.

 

[dohaku]

Imagine this announcement: a sale! (lol)

 

[lIlI]

Dreamtonics posted this, could it be a teaser?!

 

[Leon]

With SV2 still so new, I expected a sale at best. So that could be something interesting!

 

[BANKARA_TEEvie]

it loooooks like... choirrrr voicebankssss? or a feature like that of ace (an engine everyone knows the features of very very well /j) where it can do easy automated vocal-layering for a choral effect?

actually, now that i'm saying this, it does also look like it could be teasing something relating specifically to the new mixing/effects tools they added in the last update. well damn! i WILL stay tuned

 

[sapplerx]

Yeah I won't be surprised if it is another letter voicebank.

 

[BANKARA_TEEvie]

ta-daaaa! it is exactly what i thought! it seems to work a little differently from what we've come to expect from sv voicebanks, so though i kinda expected this, i *didn't* expect it to almost be like the rap mode in it having this whole unique feature-set (of which those recently added features to the engine are clearly meant to KINDA be a part of)

 

[BANKARA_TEEvie]

evil double post;
a deep look into the mind of kanru hua- i mean, the new choir voices. it seems like an actually very deep update, even if it might not be super bombastic for the average teto-using-vsynth-geek like xls was

 

[morrysillusion]

i hope people can appreciate how awesome this is and honestly how important it is for SV as a product to appeal to more people. Choir voice plug ins have always existed more than "vocal synths" as we know them, but the quality of them is so often iffy due to the fact a simple instrument VST does not contain the depth of something even like vocaloid, and especially not newer AI voice technology like SV can provide.

not only are these some really fucking good sounding choir voices but man the detail you can go into for them, is fantastic. i always wanted some good choir vocal synths to use but none of them 1) sound great 2) give the right amount of options so they dont all sound the same in style. this certainly gives a huge amount of options to make every part of the choir sound unique, and its great seeing them achieve this new vocal type after intensive research.

 

[Bookworm2]

OH MY GOD I WAS LITERALLY COMPLAINING ABOUT THIS THE OTHER DAY. This looks so cool! The demo sounds really good as well, and as a choir singer this is AMAZING. It genuinely sounds like a choir of people, not just one person with a LOT of layering. The built-in effects is a bit of a strange choice, but I guess it's for the newbies to vocal synths/music production? I dunno, but I will CERTAINLY be using them. I can't tell, does the choir effect work for non-choir vocals? I'd love to use it with SAROS, but from the demo I can't tell if it works with all voicebanks. Either way, this looks amazing, and I can't wait for it to fully release.

I can't be the only one who thinks that the Tenor voice sounds like SAROS, right?

 

[parallax_fifths]

I gave the new "unison" feature a try.

It's nice and easy to use, but does sound a bit artificial for solo voices, especially when cranked to 16-voice unison. (Although increased "spacing" helps.) However, from a quick scan of that video transcript, I think the choirs use four different voices for each of the four SATB voicebanks, so a 16-voice unison would actually be four 4-voice unisons. I dunno. After supper I'll read the transcript properly instead of scanning.

As far as solo voices go, I don't think I would use it for a lead voice, but it's already proving useful for background voices, so I'm happy.

However, I wish there were a volume control so that either outer voices could be attenuated, or perhaps vice versa. This can be sort-of achieved after-the-fact with a mid/side volume control, but it'd be nice to have more detailed control within SVS2 itself. Like a frowny-face/happy-face group volume control.

 

[parallax_fifths]

On the down side, I normally export tracks as mono. However, to make the unison feature useful, you have to export as stereo. For me this means exporting twice, once for the mono voices, and once for the stereo unison voices.

 

[BANKARA_TEEvie]

I'd love to use it with SAROS, but from the demo I can't tell if it works with all voicebanks.

if i'm not mistakennnn, i think all of these choral features are exclusive to these vocals, i'm guessing because there's a huge underlying difference with how these vocals work compared to the others... though, like rap mode and xls, i'd be surprised if it wasn't something that could become easily compatible with other vocals in the future - though i really think the way they configured these voicebanks make them the most optimal to be used in this way which ig you could say about any vb that's specifically made for a feature, but they really hammered home just how unique these guys are in how they are rendered and layered n whatever

edit: having said that, they do show off natalie in the trailer?? you can use the unison feature with any vocal then, but i think that's the one that was already in the last beta, right? either way,,, i can't say i totally get how these new vbs and the engine integration of them work yet ekhfsdfhsg

Choir voice plug ins have always existed more than "vocal synths" as we know them, but the quality of them is so often iffy due to the fact a simple instrument VST does not contain the depth of something even like vocaloid,

yeah, this is how i feel! it's not a feature that i myself could use a lot, but i've seen a ton of crossover between vocal synthesis and choir synthesis over the years (where you'd probably find that most of the non-fandom musicians using vocaloid at the time *were* using it for this very reason) and i think it's incredibly smart for an engine to finally take full advantage of this- ESPECIALLY rather than just releasing a singular choral style voicebank or something like that.
i do think that there were a lot of cases in which vocal synthesis was considered as 'just as good' if you wanted to do something choral, which meant that development in this space was really only as good as the leading singing engine itself was at the time. i'm glad to see DT realize it isn't necessarily as simple as just "use the singing synth for chorus! easy!"

gone are the days of realivox ladies, now our digital choirs can too sing with full realism *and* lyrics

 

[parallax_fifths]

if i'm not mistakennnn, i think all of these choral features are exclusive to these vocals,

The unison feature works with all SVS2 voicebanks (up to 16-voice unison), but that's it.