Other Maghni AI by VocaTone Studio and Misbah Studios

[CdrSonan]

If I recall correctly it‘s actually not a NN synth in the way the term is normally used, and they‘re just calling it AI because it‘s going to have either AI assisted or automatic tuning
I think they said during the Q&A that it‘s a concatenative engine, and then clarified what the AI is used for when answering a later question.

Either way, I completely agree with you that it‘s a huge undertaking. They announced their project way too early, and this is now making things a lot harder for them since everything happens under the eye of tons of people, so I can see why they might see the need to rush something out.
Hopefully it all works out in the end!

 

[lIlI]

Whoa whoa, slow down there, this post does not contain any major engine update news (but please check out OP for the latest monthly update)

BUT

Today Vocatone revealed Aurum and Audine's finalised redesigns, only chibi artwork available currently, but there is a preview of the full illustrations on their website. Also: they have a website. There's Oliver merch available for Ollie fans, get you a sticker.

artwork

I am liking the new details they added, also Audine is very cute.

 

[pico]

Does anyone know what they are referring to as the “pastry method” in the latest blog post?

 

[lIlI]

Does anyone know what they are referring to as the “pastry method” in the latest blog post?

No idea, I don't think it's been explained. Your best bet is probably to @ them on twitter.

 

[poser]

 

[pico]

Slightly concerned by the phrasing— is the only prototype we’re seeing the GUI prototype? As I recall, this is the component that was first programmed a long time ago. I figured that maybe the development team hadn’t had experience with the signal processing technology to create something like this, so I hope if there’s no technology prototype, there’s at least a clear plan for what existing technologies will be used to make this work to set expectations appropriately.

 

[lIlI]

It's been two and a half years since the date they originally scheduled the completion of their demo, so I really hope it's a meaty demonstration!

 

[morrysillusion]

a lot of maghni ai back when it was actively being spoken of had me very unsure of the stability of the project. next to none of it was actually development related to the engine. like i remember loads of languages being listed for support that people were getting hype about but.... none of the promises can hold up when you dont even have an engine to start working that into it. not to be a whole downer, bc id love to see them do it! but the announcement says GUI preview and 'highly-requested features' which sounds like more "we are going to tell you about all the cool things we want to do before we even do it/implement it" which is what they did before.

they definitely lacked any kind of coders adept with vocal synth technology to work on the project and if they dont have that figured out i just see the same thing repeating again. iirc they had previously hoped to get money before they had any thing coded/couldnt provide any prototype for people to trust funding them by crowd fund. i hope that this indiegogo wont be the same deal. its really important that they have some kind of prototype to tell people they can make this and imo it needs to be more than a GUI.

im not even specifically saying theyre untrustworthy in some kind of scammy way. they could 100% come out with something. but a company needs to actually act trustworthy and show theyre capable in order for them to be trusted and seen as reliable. their behavior hasnt been very reliable to me

 

[Leon]

I am tentatively excited! I think this has the potential to be really cool. And I genuinely want them to succeed. TBH if it isn't painfully expensive I'll probably back them too. I can put off Gumi for another month in the name of helping a new synth, if it looks promising. I really do hope we get a decent sneak peak though. Of course it's hard to make something substantial with a small team, but as others have said, it's been a while now and it's hard to put your money into something if you can't see the endgoal...

 

[Leon]

Here were their countdown pics. Oliver's inclusion has some people speculating about an update. I'm going to reserve judgment and just assume that they're using his image since VT created the Vocaloid, but that certainly would be a pretty hype announcement!

I'm looking forward to tomorrow either way

 

[pico]

It appears to be a new/distinct design. He has a braid on the right side of his head, the bandages don’t cover all of his face and he subjectively looks a little older? IDK.

Fingers are still crossed for engine technology demo.

 

[Leon]

Hmmmmm

 

[Mika]

The crowdfund is here!!

 

[nene]

I don't know anything about the project outside of the video... but it says it'll have 20 languages at launch?? (40 overall) :o

That sounds wild to me! I didn't realize that was a do-able launch feat?

Wishing them luck. 🫡

 

[Pup-p]

Olivers voice in the singing demo sounds a little off to me? But its early days so looking forward to seeing them improve it. Wishing them good luck on the crowd funding!

 

[Mika]

I'm quite intruiged by this part under "What Can it Do?"

Each voice has a few pre-set tonal expressions -- these are just fun parameters you can automate across the voice to change how the singer sounds, whether it be Oliver's choir voice color or his rock voice color!

Oliver rock voice color?? Interesting..

 

[lIlI]

Yay, it's so exciting to hear it at long last! The sound quality seems good, I don't hear much engine noise at all. Oliver's voice is a little wavering, though at this stage I don't know if this is the result of the engine or the intended imitation of his voice provider.

Overall, it's awesome to see a working prototype, but I am a bit disappointed it's so short and doesn't showcase Maghni's parameters/languages, or a more mainstream voicetype. Despite that, I'd love to see the project succeed. I'm glad their 20k goal seems feasible (that's less than the average salary of a single employee, surprisingly low). Having another competitor would be incredible for vocal synths, especially an engine developed with the community's ethos in mind, against unethical AI, and in favour of adding underappreciated languages simply for the love of the technology.

Good luck to them! I'm taking hints that they have more demonstrations to show down the line optimistically.

Obligatory warning that IndieGoGo does not guarantee refunds, so do donate in the knowledge that if the project goes tits up, doesn't reach its goal, or doesn't turn out as you'd hoped; your money is unlikely to be returned.

 

[Leon]

How exciting! I don't know if it's a hot take, but I kind of prefer Oliver's tone in the new one, albeit a bit shaky on some notes (mostly long notes). Assuming the stability improves, I'm happy with the demo! And I think 20k should be attainable.

I do wish Aurum and Audine had had snippets as well, and maybe a glance at some parameters? It feels a bit like going out on a limb to give them a lot of money. But I have hopes for it!

I'm a little bummed that the only tier to get all 3 physical boxes is $500 USD. I wish there was one that left out all the previous tier merch stuff to make it a little cheaper... But I'll debate with myself if I do a single bank or save up and spring for the 3

 

[lIlI]

From the silhouettes on their crowdfunding page, it's very strongly implied that Ruby, Big Al, and Sweet Ann are returning. I wonder if the latter two will keep their voice providers?

 

[pico]

Definitely happy they’ve made it to a point they’re happy with beginning crowdfunding! I do sincerely hope they succeed.

Elements of the crowdfund presentation did kind of raise my eyebrow a couple of times. Curious what users’ thoughts are on these points.

• They appear to have a functioning prototype, which is fantastic— so why not champion that more? It seems so silly to hide just a few seconds of their largest selling point within a video that otherwise doesn’t elucidate much of the product at all.
• Slightly worried about the legality and professionalism of using character designs and other products’ logos in their promotional video for their commercial product.
• As a general aside, it seems like a lost opportunity to have so much of this marketing be “fandom-centric” (OLIVER and Circus-P being ambassadors) when this is their biggest chance to open their doors to a wider audience.
• I just want more technical information. Treat your product as what it is— a technical product. “Deep learning” is effectively a buzzword in this context that tells us nothing about what your product *actually is* and how it is different from competitors. Your crowdfund is your opportunity to achieve investment. The page should be all-encompassing and shouldn’t just be used to “sell” what your product could be. It’s great that the developers want to reveal more information over livestream, but your investment pitch shouldn’t be barricaded behind something informal like that. Have it all on the table so investors can be confident and give you lots of money.
• How is supporting these languages actually going to work? When will we see their implementation? Having a concat reclist is not the same as a functioning model
• Looks like Ruby, Sweet Ann, and Big Al are coming. (Ninjad)
• TLDR glad it’s coming together but disappointed they’re still struggling to act like a company instead of a ragtag group of fandom dudes