Other Maghni AI by VocaTone Studio and Misbah Studios

[___]

Apologies for double posting.

Samples of the voice providers will be locked behind a paywall it'll be part of a tier......

so we won't be even getting that...

 

[lIlI]

Apologies for double posting.

Samples of the voice providers will be locked behind a paywall it'll be part of a tier......

so we won't be even getting that...

They've confirmed now that this isn't the case! They initially interpreted 'samples' as 'engine samples'.

 

[peaches2217]

They've confirmed now that this isn't the case! They initially interpreted 'samples' as 'engine samples'.

Ah, crap, when did they say that? I've been tab-hopping while listening so I musta tuned out

 

[cafenurse]

I agree with the points everyone has brought up, but my question is how do you pronounce Maghni?? Is it mag + ni?

mawg-knee

 

[___]

They've confirmed now that this isn't the case! They initially interpreted 'samples' as 'engine samples'.

Really? I had to leave for a minute but all they've been saying up to that point very clearly sounded like they were talking about unprocessed samples of the voice providers. Would be cool if that weren't the case.

 

[lIlI]

They have said they'll be showing us raw examples of the voice providers singing on the 7th of Jan!

 

[peaches2217]

Well, I was hoping that would clear things up and give me more confidence in the project... but nearly two hours later, I'm just even more confused. There were some neat points, but the majority of it was "We have ideas! We're very proud of it! We're gonna do so good!", followed by "Well, if you want to see any of those ideas in anything resembling context, you'll need to give us your money first ". A lot of "We're completely done, we just need the funding to make it happen!", followed by "We haven't done anything quite yet but we have some solid ideas!". A lot of exciting key phrases with little to suggest it's anything more than just words. "Can we hear samples of the synthesized vocals?" "We can show you the VPs singing some scales!". And, of course, damn-near every last bit of it locked behind a paywall.

Funding is huge and necessary, I get that. But if you have everything except the money, can we not see something? A sketch of the anticipated interface? The VPs synthesized through Utau? They're talking about all their future plans for VBs in Arabic and Portuguese and how they wanna approach universities to offer it as a learning tool for music students, but they don't even have anything but words and promises. It sounds like... well, like me when I'm talking to my girlfriend about my fanfictions. I'll talk about a huge reveal I wanna do in Chapter 24 when I haven't even finished Chapter 1.

I'm hoping and praying we'll get some kind of solid, concrete evidence of what we're investing into before the Kickstarter launches, and hell, if I like what I see of course I'll back it! But right now? If it continues like this I'm... not optimistic. And it pains me to say that after hearing them, because they sound like such fun and genuine people. But all I'm hearing is empty, empty, empty promises.

 

[lIlI]

Good news, they're going to provide some images of the GUI on the 7th (Updating the thread with relevant tidbits as they happen)

 

[peaches2217]

Good news, they're going to provide some images of the GUI on the 7th (Updating the thread with relevant tidbits as they happen)

I'm glad for that! Hhhhhhhh I just... it's like they change their answers to everything as they go along. They made a few blatant contradictions during the time I was tuned in (Discord is severely overwhelming for me so I couldn't stick around). I'm happy we are getting more, but I'm still extremely skeptical... I'll play this one by ear. After the way this has all been handled I really don't think I can trust them, but I'd still like them to succeed. I'll be happy to be proven wrong.

 

[Rolo]

Everyone’s said many things that were my concerns so I won’t say that but looking at this from a different perspective while I’m a little on the fence about the project it would be great if it succeeded. And the only way it’s going to succeed is with community help that’s why I believe it would be the most productive for voca community to just tweet their concerns at vocatone or try to talk to misbah. If I’m not mistaken someone here works with Misbah that’s a way to get your concerns to them even faster. If they’re expecting money then they should value the opinion of the community they want money from. However if they never hear these concerns because the only people who offer them feedback are people who want the project they may never know how they can do better. All this to say I think it would be good to show vocatone some of these concerns so they can know they need to do more and be more transparent if they want the community trust.

 

[cafenurse]

I was in the Q&A for a cute minute, I joined kind of late so I didn't get all of it but I was there for maybe an hour? Maybe more
I think this voice chat experience helped me better understand the kind of dynamic this company has, they're just big vocal synth fans that have really big dreams. I don't necessarily feel like ":/ this is a scam" anymore, I think my doubts have now manifested themselves into more of "I'm worried they're in way over their heads." They want to do a lot of really great, ambitious things. And it all sounds super cool if it can happen, but I worry if such a small team relying nearly entirely on crowdfunding can actually do these ambitious things. I wish them luck and I will keep my eyes on this project because if they can make their Maghni AI dreams happen, it'll be a really cool synth.


Also, if you missed the stream and were wondering why Ausgris' name was changed....they said it was because Ausgris was hard for people to pronounce and a lot of people were pronouncing it differently, plus for the Japanese fandom the katakana spelling of Ausgris was very similar to "ugly", so a lot of people were calling Ausgris "Ugly" lol. I thought that was funny. I like the name Aurum better anyway.

 

[peaches2217]

I was in the Q&A for a cute minute, I joined kind of late so I didn't get all of it but I was there for maybe an hour? Maybe more
I think this voice chat experience helped me better understand the kind of dynamic this company has, they're just big vocal synth fans that have really big dreams. I don't necessarily feel like ":/ this is a scam" anymore, I think my doubts have now manifested themselves into more of "I'm worried they're in way over their heads." They want to do a lot of really great, ambitious things. And it all sounds super cool if it can happen, but I worry if such a small team relying nearly entirely on crowdfunding can actually do these ambitious things. I wish them luck and I will keep my eyes on this project because if they can make their Maghni AI dreams happen, it'll be a really cool synth.

Yeah, I still think they're quite sketchy, but it's a lot less "They're trying to make off with our money and then drop off the face of the earth again" and more "They're in way over their heads and don't wanna let people realize that they don't have it all together". I'd be a lot less critical if they'd just... be open about that. It's clear from their contradictions and repetitions that they're scrambling and don't really have a solid, cohesive plan in place, and given how huge this project promises to be there's no shame in not knowing exactly how it's all going to work out just yet. But rather than say as much, they just went on about those plans, their plans on making more and more and more VBs ("We wanna cram as many in there as possible" I think is how they put it), their plans to get government grants to market it in an academic setting... they're waaaaaaaaaay ahead of themselves, and they would severely benefit from taking several steps back and focusing on getting the engine and their two confirmed VBs finished before going forward.

 

[___]

Time....is as valuable as money in these projects and what they should be doing is focusing more on the time time aspects than the money aspect. I don't think they're out there to do harm but also I don't think this project is at all at the stage where they have the right to be asking for money....they are just not. Something similar, smaller scale, like SynthV, took around 10 years, didn't it?

What I feel like, the approach they're trying to take, is hyping the project up to fund it to get it out as soon as possible when their approachs should be taking their sweet time with it.

 

[lIlI]

The Q&A is complete! We hope to have an audio recording published on VocaTone's channel, all tech willing.

I have typed up notes while listening, please bear in mind that I was also moderating the chat at the time, so don't take my wording as gospel, nor is this a complete summary of everything that was said. This is all paraphrasing and there is a chance I misheard. Cluemily has also provided her own transcription, feel free to cross-check between us!

Disclaimer over, let's go!

Things VocaTone has completed

Pseudocode
The software flowchart
GUI design
Phonetic dictionaries and recording lists for all languages listed in OP

Their goal is $35,000 for programming, promo, and merchandise. They have been approached with grants as alternative methods of funding.

VocaTone and Misbah Studios are two arms of one entity, one aimed at the community and one aimed at professionals outside the fandom. Misbah previously produced libraries based on rare instruments and worked with bands.

Planned features
AI to give automatic suggestions for every parameter
Physical version likely to be a USB drive rather than a CD due to the obsolescence of CD drives
At least four to five pitches
Ability to switch between XSAMPA, Arphabet, and more
Public demos when the software is ready
More centralised method of purchasing voicebanks so all are easy to buy
Develop for educational and cultural preservation purposes
Max donation tier is $5,000
Colour themes for characters, as well as light mode and dark mode
Edits made by drawing waveforms, inspired by V1's palettes
Pitch editing is similar to DeepVocal
All aspects of the phoneme can be edited
Support for Windows and Mac
Not open source
You can choose which languages you buy for a character, bundles will be discounted
Multithreaded program

Things they intend to try or are open to considering
AI suggestions for genre and emotion
Tutorials on using the software and music composition
They have contacted a vocalist with the deepest style of singing voice, and two multilingual jazz singers from Brazil and Algeria, and more
They are not looking for third party applications until the software is complete, but they are interested in working with independent groups to expand into new markets - would not need to be a company. SDK price will be public once ready.
Job plugins for microtones
Accessibility features, such as for the dyslexic and hard of hearing
Linux support
Working with other companies for collaboration or support
Voices of deceased signers based on impersonators or with permission from their estate
Different styles, like opera, idol, death metal, and country
Expand the Asian market beyond Japan and China, to places such as India

They cannot confirm
Modularity
Updates for previous Vocaloids

~Fun facts~
Aurum is based on a 1930s radio producer, inspired by the famous broadcast adaptation of War of the Worlds, and gold.
Audine is based on silver, and the novel Do Androids Dream of Electric Sheep.

When we first heard about Ausgris, they were actually working on five voicebanks for VOCALOID. Three of their names were Era, Nyx, and Weston. It was not mentioned whether they'll be resurrected on Maghni AI.

 

[Rolo]

A lot of that went over my head since I’m not a synth user but their background sounds impressive. I think as a sign of good faith they should just show these things and get some type of a test software up before they ask for that large amount. And maybe even see what other studios have going. It’s nice to have big goals but you have to be realistic with your dreams and sometimes that means compromise. There’s just no way they can do it alone if the software isn’t ready to go yet. But I guess we’ll see it sounds like they will have some stuff coming up.

 

[peaches2217]

A lot of that went over my head since I’m not a synth user but their background sounds impressive. I think as a sign of good faith they should just show these things and get some type of a test software up before they ask for that large amount. And maybe even see what other studios have going. It’s nice to have big goals but you have to be realistic with your dreams and sometimes that means compromise. There’s just no way they can do it alone if the software isn’t ready to go yet. But I guess we’ll see it sounds like they will have some stuff coming up.

That's my concern too. This is a lot, and it'd be a lot even for a large, well-established, financially-comfortable company. It's fun to daydream about or to chip away at, but they don't even have any recordings for their first two banks, much less an editor to use them in. I feel strongly that that's what they need to be focusing on right now before making promises and grand plans for anything beyond that.

 

[lIlI]

One thing they were explicit about is that they're not financially able to make any further progress on the software without further funding, so they can't create a demo or prototype unless they receive support. Unfortunately this means unless VocaTone wins the lottery, a pre-kickstarter demo of the software is off the table.

 

[peaches2217]

One thing they were explicit about is that they're not financially able to make any further progress on the software without further funding, so they can't create a demo or prototype unless they receive support. Unfortunately this means unless VocaTone wins the lottery, a pre-kickstarter demo of the software is off the table.

Precisely! I feel it's a bit (or, well, a lot) much to be talking about how much more they're going to do in the future when, presently, they can't even afford to continue with what little they have. I would feel a lot less conflicted if they centralized their focus onto what they have; right now it just feels so scattered.

 

[Maribelle]

Reading through the document the jazz singers have me intrigued tbh then I saw the comment on Sweet Ann and idk if everyone picked up on this yet but I am a Sweet Ann STAN so the non no makes me really want this to succeed cause I want hope for a Sweet Ann update ;-;

 

[___]

They've confirmed now that this isn't the case! They initially interpreted 'samples' as 'engine samples'.

They have said they'll be showing us raw examples of the voice providers singing on the 7th of Jan!

I got to chance to listen through the whole Q&A, checked their social media and let me know if I missed anything but so far...it's still sounding to me like the samples of the voice providers are behind a pay wall.



The samples being avalibe at the 7th


Wouldn't necessarily contradict the fact that it'd be behind a paywall I don't think? I dunno I'm little confused and just trying to make a sense of this. No hard feelings.