SynthV Synthesizer V Discussion Thread

[lIlI]

I'm hyped! But I hope they finish up overdue projects like the Sundial album, Saros' song contest, and his lite version; before taking on too many new things. It's been almost a year since we had an update for a few of those.

 

[Leon]

Since they were the company that dropped the first spanish voice i wonder if they'd be dropping a french voice since it's been confirmed french is in development for synthv

I think they plan to drop native Spanish VBs next (reminder that Saros's VA, who I love and who did a great job, did grow up speaking Spanish in the US, and while I'm tone deaf to it, I've heard people say they do have an accent). I do think it'd be lovely to get French added soon!!

 

[BiscuitsandTriscuits]

In my delusional heart the French support drops with Macne Nana update onto synth V fulfilling the prophecy

 

[MillyAqualine]

I want it so much to happen (and dang, I suddenly imagined a brand new outfit for Nana in case she'd get a French bank though it may not be Frenchie enough)

 

[Blue Of Mind]

Since they were the company that dropped the first spanish voice i wonder if they'd be dropping a french voice since it's been confirmed french is in development for synthv

It's also making me wonder if they're developing for SynthV 2, since it's been speculated French (and Korean) support could be locked behind a new engine.

 

[SeleDreams]

It's also making me wonder if they're developing for SynthV 2, since it's been speculated French (and Korean) support could be locked behind a new engine.

To me it's almost guaranteed that synthv 2 is the target

 

[Leon]

READY FOR MY NEXT FEAST, ECLIPSED

 

[morrysillusion]

i was intrigued from the start of course, but the sample at the end? i have never been so hyped in my life for a new voice

 

[Heruru Meruru]

We're getting a metal screaming vocal?!

 

[Leon]

OPERA REVIVAL TOO

 

[pico]

Cannot WAIT for more genre specified vocals for this engine... long overdue. I think Eclipsed will do a great job.

Hoping very very very very much that these fellows have good character designs. Too early to pass any judgement but the videos make me wonder what's going on there. I think it's interesting how far these videos diverge in presentation from the Starry Court, though. Shifting to a more explicit fandom focus? Not what I expected, at all.

 

[Chuchu]

Oh, can't wait for the full demos, but I'm really excited for the metal voice ghhffdaff I wonder which specific genres the rest of the tone chroma series will cover Maybe EDM in the direction of hardcore/rave, breakbeat? I'd love a voice similar to Keith Flint tbh.

 

[Blue Of Mind]

Ninezero has a brother, and Prima's back! /joking

But seriously, I'm glad Eclipsed Sounds are taking risks by creating more genre specific VBs for SynthV. Screamo VBs for commercial vocal synths have been begged by the fandom since the V2 days, when producers figured out how to get the Cryptonloids (and Gumi and Gakupo) to scream through external programs.

 

[Chuchu]

Could there be a possibility that Eclipse might create multiple voicebanks for a colour? R-01 and V-01 makes me wander if there will be a R-02, V-02, etc.

It would be awesome if there were multiple operatic voicebanks to create a Gregorian chant

 

[SeleDreams]

Ninezero has a brother, and Prima's back! /joking

But seriously, I'm glad Eclipsed Sounds are taking risks by creating more genre specific VBs for SynthV. Screamo VBs for commercial vocal synths have been begged by the fandom since the V2 days, when producers figured out how to get the Cryptonloids (and Gumi and Gakupo) to scream through external programs.

People assume it's screamo vb but i'm thinking it might be part of SynthV 2's new features.
vocal modes that allow to produce screamo vocals and more opera vocals in general, the same way as they implemented the rap vocals feature.

 

[morrysillusion]

if it were a program wide feature i would be even more happy. however i feel for something like this, some serious training data from the original VP would really be needed to make it sound good? i am thinking about how V4, some voices used the same growl while others had unique ones, and it often made a big difference. itd be great to have this feature for all voices but it seems like it'd be tough to make it sound good unless the bank has dedicated information to train off of (but i am only taking guesses on how this kinda stuff works). and if that were the case, i'd guess R-01 would probably be an example of having that data. but it seems more believable to me that R-01 just has its own vocal mode unique to itself (and that new engine changes for synth v 2 make this kind of thing possible)...

 

[SeleDreams]

Since they were able to introduce cross language synthesis to make voices sing phonemes they never recorded through common training data, where voicebanks use data collected from training other voicebanks to have added features, I think it is doable

But we'll see once we get more informations.

 

[lIlI]

As this is a notable release, I've made a separate thread. That makes the discussion easier to find when people search for it!

 

[SeleDreams]

Well, the live was pretty underwhelming. no progress on the korean and french language, and the jump in quality didn't feel that crazy.
It's mostly some ease of use improvements but nothing too major

 

[inactive]

So there I was, reading an article on Ars Technica, when an advertisement for an online casino appeared off to one side. Not unsurprisingly, this ad featured a slop-generated image. But something seemed awfully familiar about this particular slop image:

Now that's an image that will make SynthV users go "hmm..."