Do you know what the entire thing here reminds me of? Yup, the "Stella" vocal synth fiasco.
I'm a bit surprised I'm not seeing more talk (in general, not just here) about the primary purpose of the crowdfund - to help a friend out. This is genuinely such a crazy thing to just casually drop like that and I would feel incredibly betrayed if I were a backer. Is this not actual fraud?
I'm not expecting a lawsuit to come of this or anything, since technically yes the money did need to go to their single developer who wouldn't be able to do any work if they ended up being deported, but people were not contributing to the crowdfund knowing that was happening. That's actually insane to me.
I'm not expecting a lawsuit to come of this or anything, since technically yes the money did need to go to their single developer who wouldn't be able to do any work if they ended up being deported, but people were not contributing to the crowdfund knowing that was happening. That's actually insane to me.
Wait, I completely skipped over that part. What?? How do they expect anyone to trust them after admitting that....
At least with the Stella situation, Planty was just planning on releasing albums. In a reality where she wasn't revealed to be fake, I can't even see the albums raising that much money. I am noticing the similarities since you brought it up. This one is just more damaging than Stella's situation.
The revelations paint a picture of a project that unfortunately failed at every level. I believe the folks saying that this project was draining, damaging, even traumatizing. I feel bad for them and I hope they can create space for themselves from this for the future. Simultaneously, I'm frankly aghast at the lack of accountability from basically all of these people. There's too much that could be said about too many things in that area so I won't go down that road.
I have some thoughts on the continuation of this project that some folks seem to be committed to.
In my opinion, they should dissolve Crescendia (especially if Crescendia is Natalia's entity?), and formally cancel Maghni.
Good for them if these guys want to develop an engine on their own time, but it makes zero sense to do so on this fundamentally broken foundation. "Maghni" as it was promised on social media and in the IndieGoGo campaign ostensibly does not exist. It is still not clear if there are plans to make it exist in the specific ways that were promised. I personally doubt it. So, can it.
I think it may be tempting for the remaining contributors to think they can "make good" on the Maghni promises by spinning up an engine & maybe reusing the Maghni frontend. In my opinion, unless these folks are contractually obligated to do so (which I seriously doubt, if it is true that employment terms were discussed on the phone and not committed to paper, and if the NDA permits this much to be said) it is a bad idea to accept that baggage, no matter how well-meaning. And, clearly, the reality of a complete engine of any kind is very far away.
To be clear, I do believe that these folks mean well, but I think this is entirely the wrong way to go about it at this point. In my opinion, they can't be specific about deliverables because they do not know what is within their capacity, and on what timelines. That being the situation, it may be the case that some or all of this may never be within their capacity. Even unintentionally, the public is still effectively still being led on. They can choose to stop doing that. If they're valuable, they can take their prototypes, do something else, and come back when it exists.
Maghni is a failed business venture. The quicker they accept this, the easier this becomes for everyone, especially their fans that were led astray.
My favorite excerpt from the blog post:
EDIT: Unproductive comments from me tonight but wow I am watching posts get made on Twitter right now and these posts are making things worse. Oh my god, every post makes the situation worse. Stop posting.
I have some thoughts on the continuation of this project that some folks seem to be committed to.
In my opinion, they should dissolve Crescendia (especially if Crescendia is Natalia's entity?), and formally cancel Maghni.
Good for them if these guys want to develop an engine on their own time, but it makes zero sense to do so on this fundamentally broken foundation. "Maghni" as it was promised on social media and in the IndieGoGo campaign ostensibly does not exist. It is still not clear if there are plans to make it exist in the specific ways that were promised. I personally doubt it. So, can it.
I think it may be tempting for the remaining contributors to think they can "make good" on the Maghni promises by spinning up an engine & maybe reusing the Maghni frontend. In my opinion, unless these folks are contractually obligated to do so (which I seriously doubt, if it is true that employment terms were discussed on the phone and not committed to paper, and if the NDA permits this much to be said) it is a bad idea to accept that baggage, no matter how well-meaning. And, clearly, the reality of a complete engine of any kind is very far away.
To be clear, I do believe that these folks mean well, but I think this is entirely the wrong way to go about it at this point. In my opinion, they can't be specific about deliverables because they do not know what is within their capacity, and on what timelines. That being the situation, it may be the case that some or all of this may never be within their capacity. Even unintentionally, the public is still effectively still being led on. They can choose to stop doing that. If they're valuable, they can take their prototypes, do something else, and come back when it exists.
Maghni is a failed business venture. The quicker they accept this, the easier this becomes for everyone, especially their fans that were led astray.
My favorite excerpt from the blog post:
This shit is so unserious.
EDIT: Unproductive comments from me tonight but wow I am watching posts get made on Twitter right now and these posts are making things worse. Oh my god, every post makes the situation worse. Stop posting.
Last edited:
I wonder if part of them feels an obligation to finish this project, especially with how everyone was banking on the fact the PFX engloids got discontinued and were going to be on Magnhi (at least Oliver and Ruby were confirmed, I can't remember if Big Al and Sweet Ann were 100% confirmed. I just vaguely remember that people mentioned them in connection to Ruby being announced for it).
I don't think anyone was being particularly malicious, like what Pico said. But they should just count their losses at this point. I can't imagine people will trust them if the project did end up being finished, and they'd lose more money on that alone. I try to be sympathetic as much as I can be to projects with small dev teams, but come on.
yes because that is the most pressing matter of the whole situation.
I don't think anyone was being particularly malicious, like what Pico said. But they should just count their losses at this point. I can't imagine people will trust them if the project did end up being finished, and they'd lose more money on that alone. I try to be sympathetic as much as I can be to projects with small dev teams, but come on.
yes because that is the most pressing matter of the whole situation.
Well... here we are. I can't say I expected things to blow up quite like this. Right now I'm withholding any kind of opinion on who's to blame because nothing's all that clear and long explanation documents attached to twitter posts aren't exactly known for being the most reliable sources anyway. It seems like a lot of the devs were victims in this situation, though, and I wish them luck and hope they can move forward and continue to pursue ambitious creative endeavors with success in their futures. That being said, I also have to agree that pushing to release an engine with the scraps of Maghni that do exist is probably a bad idea.
The history and state of Maghni that we do know about from the devs' letters make it seem like the engine has miles to go before anything remotely presentable is possible. I wonder how much documentation actually exists and in what state the code for the engine actually is right now. We know from Cocoa's posts that the raw output at the time of the Oliver demos was... well, honestly so bad it literally sounds like a horror movie, and the development team has been embroiled in setback after setback for months now. Especially with a gutted team left to work on the engine and essentially no money to build it, I doubt we'll see or hear anything for years. This is life support, and it might just be better to let it go. How can you move forward with something that's a passion project at its core when something like this goes down?
And, regarding the demos (and, really, all of the devs' responses at this point): I don't think the attempts at damage control are doing anything but making this worse. The issue about them using SV isn't that they used SV. It's that they used it to falsely promote a product. There was no timing model! The consonants sounded terrifying! The output they presented was not actually made by the engine they claimed it was made by! And, to make matters worse, they were using those demos to crowdfund the fix for the visa issue while claiming it was going to the engine development. That is a textbook definition of a scam! In trying to defend yourself against a claim of using Synth V, you've made it irrefutible that you did, in fact, scam your backers. There's no way to regain trust after that, even if the intent wasn't malicious.
Nothing about this is being handled well. This is bigger than a twitter fight filled with long Google docs of receipts. Accountability is being dodged left and right. I have sympathy for those who are trying to make something out of this wreckage, but I have to admit the red flags were there. Slap LLCs on things all you want, but at the end of the day, this really does feel like yet another ambitious fan project that ultimately failed because those behind it bit off way more than they could chew.
New Twitter post is here :
x.com
Basically none of the singing audio outputs were generated from Synth V, they just use Synth V to draw the pitch curve, then copy to Maghni through their proprietary tooling.
They did admit to utilize RVC in one portion of a song to generate a speaking voice for Oliver and spliced some consonants from Oliver's voice provider's recordings into the Maghni AI output.
Actual raw outputs from Maghni AI have been revealed:
cocoa 👽🍫 (@ccodaze) on X
Hello everyone! I'm here to provide some samples of audio from early Maghni AI renders. I have seen rumors going around that they were generated with Synthesizer V, so I wanted to clear that up.
Basically none of the singing audio outputs were generated from Synth V, they just use Synth V to draw the pitch curve, then copy to Maghni through their proprietary tooling.
They did admit to utilize RVC in one portion of a song to generate a speaking voice for Oliver and spliced some consonants from Oliver's voice provider's recordings into the Maghni AI output.
Actual raw outputs from Maghni AI have been revealed:
Last edited:
Hard to keep up with but it was clarified that SynthV output was used as the inference for RVC that they later spliced with the Maghni output.
Regardless of what was spliced from where on what demos (everyone involved in the convo seems to agree that at least some misrepresentation took place), I just can't get over this statement. It's as if the software they developed was a mysterious 3rd party that put them into an uncomfortable position they had no control over. Nobody was forcing you guys to misrepresent your output in any way to the public. Hint: the renders you shared should have been awful, if that's what the engine was.
I think they practically realize this, but it's like this post betrays a "it can't be helped" or "you gotta do what you gotta do" attitude that seems to have been the root of so much of the dishonesty. Nobody was forcing any of these people to do anything, beyond deliver on what was paid for. The deception was optional.
Eggtan's demo is "Right Where You Are". They composed the song.
Honestly, at this point, even if maghni were to come out, its reputation is so destroyed that i doubt most people of the community would even want to financially support it by buying it anymore
Now that it is clear that based on my understanding, 6/7 of all demos presented (other than the final demo "Right Where You Are") are the truthful raw output of Maghni AI, where as "Right Where You Are" (the Youtube video below) might involve RVC splicing.
Last edited:
Especially when the communication was not being delivered actively from time to time, as it was done in the past. I think if let say they encounter any challenges or problems, they should be transparent from the earliest stage of Maghni AI synth development, or they should have just establish their discord or forum site so that they can share technical updates on the voicebank with sponsors, and perhaps people who know enough about the vocal synth can give some constructive input so that they know which way the team should take. Do not underestimate sponsors or any other users/fans who were enthusiastic with the project, because they might be the ones that could help out even with the technical stuffs.
With the lack of communication, everything could be spiralling in a more negative way. The staff involved in the development will suffer mentally and physically, while the sponsors will suffer from waiting their promise to be delivered. This is not healthy for everyone.
I feel like maybe this is a reasonable posture to take for non-profit or open-source projects, but this was (and is-- these folks are still talking about protecting their IP for a commercial release) a for-profit and closed-source venture. Mining the public for free contributions would have made the current situation look even worse.
At this point, they should just cancel Maghni Ai, I don't know. This is already enough of a PR disaster as is. As much as I'd love to see Oliver again, I think it's better that they just...I don't know, not bother anymore. Nothing can save this project at this point with all this coming out.
I mean, sure, they could insist on for-profit and closed-source venture, if only they are capable or producing alpha/beta MVP of Maghni AI within these few days or weeks. However, we can see how things are going on currently, and keep on procrastinating is bad for their health too. At the current state, things need to end so that no more time is wasted and everyone could move on.
For oliver at this point they might as well try to reuse the recorded data to make a synthv bank of him
the two key disasters of this two me are regarding their use of money (the kickstarter basically being made to get money to stop someone from being deported and it was clear that was like, the entire priority over making a software) and truth of what they promise (faking demos, to whatever degree that is). incredibly significant errors to make that i see no way to come back from. who knows how theyd use more money? how will we know they arent faking demo quality? it was already growing to be a big waste of time with little progress and drawing things out and not delivering rewards etc... these new things just put a nail in the coffin to me. i hope they know whats best for them and simply put the project down. none of these people should be leading a project like this imo, and its reputation isn't gonna get wiped away very easily.