short demo with Oliver's in progress voicebank
i hate to say it but, maybe also a fault of mixing too, i couldnt understand most of anything that was being said here... i kinda had that trouble with prev demos but i also think i was able to read lyrics alongside them too which helps. this one i couldnt understand well at all and there arent lyrics to read. i think this is a bit worse on his low range
i hate to say it but, maybe also a fault of mixing too, i couldnt understand most of anything that was being said here... i kinda had that trouble with prev demos but i also think i was able to read lyrics alongside them too which helps. this one i couldnt understand well at all and there arent lyrics to read. i think this is a bit worse on his low range
V3 Oliver was quiet as a VB as well, but I couldn't understand the lyrics in the demo either. I'm hoping this is just an issue with his beta bank or a quirk with the tuning in the demo.
i dont have time to read all of this right now, but seems important to put here.
when i read all this myself i may return to add my thoughts lol.
EDIT:
you know ive always been iffy about how vocatone has handled their development of/work of maghni ai. and while its good theyve made this choice after seeing issues its just.... idk. its so weird to me how this development has gone. a long time vocal synth company that has felt... idk, incapable of managing a vocal synth project? what exactly has been an issue for them to make new vocal synth banks as a company that has before?? i know making a bank is different than a whole engine, but they did go into this hard with the intention to make an engine, and their experience spoke to that seeming to be done with confidence (along with the confidence of many supporters who know their distory in creating vocal synth related products). maybe im saying that in a very harsh way. but the flow of events in these last years has just left an impression of vocatone on me, not one id expect from a company with the prior history it has in the vocsynth space. not to say this whole project is a mess or something, its just odd that it has gone the way it has.
additionally, odd to hear about aurum's voice changing, thats kind of a shame and im not sure how to feel about that for people who already preordered them, while also having no sample of them right now either.
VocaTone Studio
VocaTone Studio is pioneering the next generation of voice synthesizer AI software.
www.vocatone.studio
when i read all this myself i may return to add my thoughts lol.
EDIT:
you know ive always been iffy about how vocatone has handled their development of/work of maghni ai. and while its good theyve made this choice after seeing issues its just.... idk. its so weird to me how this development has gone. a long time vocal synth company that has felt... idk, incapable of managing a vocal synth project? what exactly has been an issue for them to make new vocal synth banks as a company that has before?? i know making a bank is different than a whole engine, but they did go into this hard with the intention to make an engine, and their experience spoke to that seeming to be done with confidence (along with the confidence of many supporters who know their distory in creating vocal synth related products). maybe im saying that in a very harsh way. but the flow of events in these last years has just left an impression of vocatone on me, not one id expect from a company with the prior history it has in the vocsynth space. not to say this whole project is a mess or something, its just odd that it has gone the way it has.
additionally, odd to hear about aurum's voice changing, thats kind of a shame and im not sure how to feel about that for people who already preordered them, while also having no sample of them right now either.
Last edited:
I almost forgot about Maghni Ai to be honest. What I’m getting from that post is that there’s a new company in charge of the software development and VocaTone is shifting to just developing voices for the software?
Not surprised about the Aurum news, I cannot recall for the life of me if this rumor can be attributed anywhere but I remember reading, somewhere with regard to information pre and early Maghni, that Aurum's VP wasn't a professionally trained singer. That's never an automatic out for producing a professional synth, but it could become a disqualification depending on the individual's personal capabilities. Based on the verbiage of the announcement... perhaps the VP is part of the reason that the VB was delayed in addition to the self-admitted lack of experience with this sort of development.
The article is at least certainly right that the impression given by the IndieGoGo campaign and promotional materials is one where VocaTone is an extremely significant piece of the pie, particularly when "from the developers of Oliver" was used as the big chip in the marketing game to fans. "Don't you like Vocaloid Oliver? You'll LOVE this!" The implication is that the engine is coming FROM these people as well as the VBs. So I think it makes sense from the perspective of an engine developer to say, "no, we are the engine developers, and here is a 3rd party developing flagship vocals for our software that we are investing our time in for the benefit of both parties." This certainly did not stand out when VocaTone were the ones handling promotion and the frontend of the IndieGoGo, not Misbah.
I was confused by the quote "would like to formally distance ourselves from direct work on Maghni AI and its libraries" because I cannot interpret if this means they are no longer developing Aurum, Oliver, and Audine at all outside of character stuff or if this is poorly phrased and they mean "libraries in general outside of the libraries we are involved with". If it is the former, that is really poorly communicated because the talk about the new VP continues to imply that they are involved with the development.
I think my last point of surprise is the appearance of this new character, Cinna, who I do not recall appearing in any of the discussions or q&a's, but I may be forgetting or perhaps they entered the picture later on.
Perhaps in retrospect this announcement will be a big point of turnaround as far as public perception on the engine demos and the corporate hijinks, hoping for all good things for all involved parties in the future.
The article is at least certainly right that the impression given by the IndieGoGo campaign and promotional materials is one where VocaTone is an extremely significant piece of the pie, particularly when "from the developers of Oliver" was used as the big chip in the marketing game to fans. "Don't you like Vocaloid Oliver? You'll LOVE this!" The implication is that the engine is coming FROM these people as well as the VBs. So I think it makes sense from the perspective of an engine developer to say, "no, we are the engine developers, and here is a 3rd party developing flagship vocals for our software that we are investing our time in for the benefit of both parties." This certainly did not stand out when VocaTone were the ones handling promotion and the frontend of the IndieGoGo, not Misbah.
I was confused by the quote "would like to formally distance ourselves from direct work on Maghni AI and its libraries" because I cannot interpret if this means they are no longer developing Aurum, Oliver, and Audine at all outside of character stuff or if this is poorly phrased and they mean "libraries in general outside of the libraries we are involved with". If it is the former, that is really poorly communicated because the talk about the new VP continues to imply that they are involved with the development.
I think my last point of surprise is the appearance of this new character, Cinna, who I do not recall appearing in any of the discussions or q&a's, but I may be forgetting or perhaps they entered the picture later on.
Perhaps in retrospect this announcement will be a big point of turnaround as far as public perception on the engine demos and the corporate hijinks, hoping for all good things for all involved parties in the future.
Last edited:
Maghni has revealed a collab with Virvox to bring Kotarou to their engine. I hope they haven't announced this too early, as Maghni AI still appears to be in the throes of development, and already has a decently sized launch lineup on their plate. Fingers crossed fans don't have to wait too long to get ahold of his voice.
Last edited:
I even noticed that they're planning to release not only Kotarou but all of the Virvox characters to Maghni AI.
It seems like it may be a good step forward for Maghni to be more firmly under Misbah -- or, well, Crescendia LLC I guess. VocaTone being kind of the face of them as a previously established company wasn't a bad idea to start, but it does kind of muddy the water. I'm curious to see how it goes.
I wonder how Aurum's new VP will sound. We got the VP sample and then an Aurum alpha singing sample a while back, but iirc both of these were very late days or post-official crowdfund ending, so it's not like people had one idea for what they were buying and then had it changed out from under them. Who knows, maybe this new Aurum VP will be even better :)
I am also a little dubious of announcing a collab before the engine is even released, but perhaps it's a good thing. "Hey look, another company that's already released vocals wants to work with us" type deal.
I wonder how Aurum's new VP will sound. We got the VP sample and then an Aurum alpha singing sample a while back, but iirc both of these were very late days or post-official crowdfund ending, so it's not like people had one idea for what they were buying and then had it changed out from under them. Who knows, maybe this new Aurum VP will be even better :)
I am also a little dubious of announcing a collab before the engine is even released, but perhaps it's a good thing. "Hey look, another company that's already released vocals wants to work with us" type deal.
Hi, I can clear this part up a bit!
While VocaTone has had input in selecting the new VP, as it is important to their character vision, and they continue to be the driving force behind the illustrations and other aspects, they are not involved directly within the development of the libraries themselves - they haven't been for quite some time actually. (Briefly touched upon in the statement under the wording of "handing off their development to members of the Maghni team," as well as in the section of "four library developers who have been helping us with Audine, Aurum, etc."
Unfortunately, all three libraries had to be contracted out to finish their development(s), of which I was brought on for Audine in March-April, and the other developer for Aurum was brought on slightly before or around the same time. VocaTone has a great knack for concepts and character driven developments, but as mentioned in the statement they overestimated their abilities. While they made the right decision, the timing in which we were onboarded has made it increasingly difficult to achieve the initial projected release(s), especially in the case of Aurum. [While I enjoyed his previous VA's tone, it truly can not be overstated how much of a mess his voicebank was and how many hours of development were essentially wasted as the results simply could not be brought up to par with the other vocals without a re-record / complete overhaul T_T)
His new VA has some old DiffSinger samples! While it's not an exact proof of concept, it does give a great general idea.
Those clips are really nice! I don't want to be harsh, but I like it better than the kind of "cursive" singing thing the old VP clip had going on. I'm excited to see the result, although I know it will take a while now :')
There will still be an indie/cursive focused mode I believe, for those who wanted to see that voice type!
So, we are approaching the end of December. Since we're approaching the end of the year, it seemed like an appropriate time to check in on the ongoing vocal synth projects. Here are my observations.
1) (Almost) All of the verifiable milestones provided for 2024 were missed.
The last formal update to the project was the "Maghni Moment 1" video shared on the IndieGoGo updates page in May. In this video, (t=283) they outlined the expected milestones for the project, and stated that the project was aiming for a Q1 2025 release.
(direct link in case that pic isn't rendering... vocaverse dead )
The May 2024 publicly verifiable goal of "Beta demo #1" was the "clairo / amoeba" demo discussed in this thread (pg 17). Afterward, nothing public-facing came to be. "Beta demo #2, "Editor in action", "Duet", "Audine demo", and "Donovan demo" are nowhere to be seen.
2) None of the recent goings-on of the project have been communicated via the IndieGoGo.
I think this can probably be explained at least partially by the "VocaTone/Misbah" split, since it seems like VocaTone were the ones primarily handling the crowdfund and communications. Austin, VocaTone head, was the one posting there.
Since October, the primary source for information looks to be the Twitter account @CrescendiaEng, which shared the new homepage and retweeted a clip of the GUI that month. This makes sense given the presence of the new LLC and etc, but since this change of information delivery was not communicated to backers via the IndieGoGo campaign, I'm not sure how you'd get this information to them. Assuming people will get it via algorithmic proximity might be safe based on the audience, but feels like less sure of a bet especially given recent events.
Going "stealth mode" seems to be somewhat in the project's favor at this point, but these transitions should definitely be made clear so no-one ends up in trouble.
3) Engine delayed & implied eventual release will be "early access".
So that brings us to today. @CrescendiaEng acount retweets this thread written by Cinna. To summarize: ackgnowledgement of silence, reward shipments, "determined to begin early access and launch commercially in 2025". Note that they do not use the word "delay" (I think they should) but they have changed the goalpost from "Q1 2025" to "sometime in 2025". So essentially the same thing.
Laying all of this out in a list almost feels wrong, or in the vein of "kicking them while they're down". But, I think these statuses deserve to be clearly laid out given the context of the project, when the developers are not doing it themselves.
About a year ago, posts and discussion about the IndieGoGo in this thread speculated that this project would suffer delays and observed patterns matching comparable past projects with management, project scope, and presentation. A lot of folks thought this was unnecessarily harsh or preemptive discussion. I truly can understand that perspective. It was negative. But ultimately, we are here now... so what comes next?
So despite this year essentially being summed up as a series of "bummers" for this project, there are for sure some upsides. For one, it does seem like they still have fuel in the tank.
The phrasing of the Twitter thread inspires some confidence... it's the first time I felt that the developers were speaking on truly achievable timelines. Of course, with regard to the engine, since we know so little, it's really hard to say... but I certainly think that 2yrs of development time is far more actionable than one, even if 2 is also very aspirational for their stated scope. It is a little interesting that they keep soft-announcing new voicebanks without the engine's appearance, but hard to say what all will go into that.
I also think deciding to commit to sending out the backer rewards is a VERY smart choice for them as kickstarter and indiegogo backer awards are famously complex and expensive aspects of these projects. Absolutely a good choice to get that out of the way and at least bring something to their investors.
Maghni AI delay is like fork in kitchen but ultimately it looks like they have taken the steps necessary to position themselves advantageously for 2025. It would have been nice to see it happen from the jump, but progress is progress. Fingers crossed for more good news for them.
1) (Almost) All of the verifiable milestones provided for 2024 were missed.
The last formal update to the project was the "Maghni Moment 1" video shared on the IndieGoGo updates page in May. In this video, (t=283) they outlined the expected milestones for the project, and stated that the project was aiming for a Q1 2025 release.
(direct link in case that pic isn't rendering... vocaverse dead )
The May 2024 publicly verifiable goal of "Beta demo #1" was the "clairo / amoeba" demo discussed in this thread (pg 17). Afterward, nothing public-facing came to be. "Beta demo #2, "Editor in action", "Duet", "Audine demo", and "Donovan demo" are nowhere to be seen.
2) None of the recent goings-on of the project have been communicated via the IndieGoGo.
I think this can probably be explained at least partially by the "VocaTone/Misbah" split, since it seems like VocaTone were the ones primarily handling the crowdfund and communications. Austin, VocaTone head, was the one posting there.
Since October, the primary source for information looks to be the Twitter account @CrescendiaEng, which shared the new homepage and retweeted a clip of the GUI that month. This makes sense given the presence of the new LLC and etc, but since this change of information delivery was not communicated to backers via the IndieGoGo campaign, I'm not sure how you'd get this information to them. Assuming people will get it via algorithmic proximity might be safe based on the audience, but feels like less sure of a bet especially given recent events.
Going "stealth mode" seems to be somewhat in the project's favor at this point, but these transitions should definitely be made clear so no-one ends up in trouble.
3) Engine delayed & implied eventual release will be "early access".
So that brings us to today. @CrescendiaEng acount retweets this thread written by Cinna. To summarize: ackgnowledgement of silence, reward shipments, "determined to begin early access and launch commercially in 2025". Note that they do not use the word "delay" (I think they should) but they have changed the goalpost from "Q1 2025" to "sometime in 2025". So essentially the same thing.
Laying all of this out in a list almost feels wrong, or in the vein of "kicking them while they're down". But, I think these statuses deserve to be clearly laid out given the context of the project, when the developers are not doing it themselves.
About a year ago, posts and discussion about the IndieGoGo in this thread speculated that this project would suffer delays and observed patterns matching comparable past projects with management, project scope, and presentation. A lot of folks thought this was unnecessarily harsh or preemptive discussion. I truly can understand that perspective. It was negative. But ultimately, we are here now... so what comes next?
So despite this year essentially being summed up as a series of "bummers" for this project, there are for sure some upsides. For one, it does seem like they still have fuel in the tank.
The phrasing of the Twitter thread inspires some confidence... it's the first time I felt that the developers were speaking on truly achievable timelines. Of course, with regard to the engine, since we know so little, it's really hard to say... but I certainly think that 2yrs of development time is far more actionable than one, even if 2 is also very aspirational for their stated scope. It is a little interesting that they keep soft-announcing new voicebanks without the engine's appearance, but hard to say what all will go into that.
I also think deciding to commit to sending out the backer rewards is a VERY smart choice for them as kickstarter and indiegogo backer awards are famously complex and expensive aspects of these projects. Absolutely a good choice to get that out of the way and at least bring something to their investors.
Maghni AI delay is like fork in kitchen but ultimately it looks like they have taken the steps necessary to position themselves advantageously for 2025. It would have been nice to see it happen from the jump, but progress is progress. Fingers crossed for more good news for them.
This is definitely not a good sign. Let's hope it doesn't end up like Friday night Funkin...
Frustratingly, Imgur embeds don't work anymore due to an update on Imgur's end.So, we are approaching the end of December. Since we're approaching the end of the year, it seemed like an appropriate time to check in on the ongoing vocal synth projects. Here are my observations.
1) (Almost) All of the verifiable milestones provided for 2024 were missed.
The last formal update to the project was the "Maghni Moment 1" video shared on the IndieGoGo updates page in May. In this video, (t=283) they outlined the expected milestones for the project, and stated that the project was aiming for a Q1 2025 release.
(direct link in case that pic isn't rendering... vocaverse dead )
The May 2024 publicly verifiable goal of "Beta demo #1" was the "clairo / amoeba" demo discussed in this thread (pg 17). Afterward, nothing public-facing came to be. "Beta demo #2, "Editor in action", "Duet", "Audine demo", and "Donovan demo" are nowhere to be seen.
2) None of the recent goings-on of the project have been communicated via the IndieGoGo.
I think this can probably be explained at least partially by the "VocaTone/Misbah" split, since it seems like VocaTone were the ones primarily handling the crowdfund and communications. Austin, VocaTone head, was the one posting there.
Since October, the primary source for information looks to be the Twitter account @CrescendiaEng, which shared the new homepage and retweeted a clip of the GUI that month. This makes sense given the presence of the new LLC and etc, but since this change of information delivery was not communicated to backers via the IndieGoGo campaign, I'm not sure how you'd get this information to them. Assuming people will get it via algorithmic proximity might be safe based on the audience, but feels like less sure of a bet especially given recent events.
Going "stealth mode" seems to be somewhat in the project's favor at this point, but these transitions should definitely be made clear so no-one ends up in trouble.
3) Engine delayed & implied eventual release will be "early access".
So that brings us to today. @CrescendiaEng acount retweets this thread written by Cinna. To summarize: ackgnowledgement of silence, reward shipments, "determined to begin early access and launch commercially in 2025". Note that they do not use the word "delay" (I think they should) but they have changed the goalpost from "Q1 2025" to "sometime in 2025". So essentially the same thing.
Laying all of this out in a list almost feels wrong, or in the vein of "kicking them while they're down". But, I think these statuses deserve to be clearly laid out given the context of the project, when the developers are not doing it themselves.
About a year ago, posts and discussion about the IndieGoGo in this thread speculated that this project would suffer delays and observed patterns matching comparable past projects with management, project scope, and presentation. A lot of folks thought this was unnecessarily harsh or preemptive discussion. I truly can understand that perspective. It was negative. But ultimately, we are here now... so what comes next?
So despite this year essentially being summed up as a series of "bummers" for this project, there are for sure some upsides. For one, it does seem like they still have fuel in the tank.
The phrasing of the Twitter thread inspires some confidence... it's the first time I felt that the developers were speaking on truly achievable timelines. Of course, with regard to the engine, since we know so little, it's really hard to say... but I certainly think that 2yrs of development time is far more actionable than one, even if 2 is also very aspirational for their stated scope. It is a little interesting that they keep soft-announcing new voicebanks without the engine's appearance, but hard to say what all will go into that.
I also think deciding to commit to sending out the backer rewards is a VERY smart choice for them as kickstarter and indiegogo backer awards are famously complex and expensive aspects of these projects. Absolutely a good choice to get that out of the way and at least bring something to their investors.
Maghni AI delay is like fork in kitchen but ultimately it looks like they have taken the steps necessary to position themselves advantageously for 2025. It would have been nice to see it happen from the jump, but progress is progress. Fingers crossed for more good news for them.
On topic: this is a good post, I appreciate you keeping track. I've also been documenting Maghni's progress from the beginning, and one strong criticism I have is their failure to communicate transparently: mainly acknowledging when they miss a deadline. Multiple times they've promised regular updates within a specific timeframe, then gone silent for months (in some cases years) without clarification. If they forgot their promises, that implies a serious lack of organisation behind the scenes, one which has made me wary of investing in the project.*
Likewise, their methods of communicating are sporadic with little consistency on where they'll come from, making it very hard for backers to track progress. The personal account of a dev, instead of an official blog post or announcement from a company account, doesn't feel like the best choice.
Although I appreciate the inevitability of growing pains with such a big undertaking, Maghni has so far had nothing but growing pains, and they're now almost
A whole year's window suggests great upheaval behind the scenes, 2025 will be an important time for this project. The good part is that a big window combined with no specific promises suggests they've learned and acknowledged...internally...that hitting specific deadlines isn't possible for them at this point in development.*The other possibility is that they hope if they don't acknowledge they missed a deadline, they'll avoid backlash because people will forget about it. While this probably works for most fans, I remember Maghni...I remember them all...
Last edited:
Right. I'm completely on the same page with you here, I think. The complete opacity of the project and its communication has not worked to its' benefit.
The statement has historically been that public transparency is a fine line to walk for a commercial project, but ultimately, they chose to pursue public, not private, investment. Making the public your investors ultimately means ceding some of that secrecy as a result of that decision. As it stands right now, they have not been treating those IndieGoGo backers well.
Like you mention, it does seem like quite a bit happened internally to cause this cascade. I'm sure much of it is technical. Once again, the modality of the product they are reaching for is brand-new and will not be a walk in the park for development and implementation. I agree that the ultimate impression of the new release "range" is that they have finally reckoned with this. But I'm also sure that much of it is also related to people. The VocaTone blog post reveals enough of that on its own.
I truly, truly hope... that this year, they'll be able to use this new corporate structure to truly do what they intend to do with this project, communicate with their investors, and stop beating around the bush. It can only benefit them to keep it real. And yes, that they loop everyone in on an "official", set form of communication.
The statement has historically been that public transparency is a fine line to walk for a commercial project, but ultimately, they chose to pursue public, not private, investment. Making the public your investors ultimately means ceding some of that secrecy as a result of that decision. As it stands right now, they have not been treating those IndieGoGo backers well.
Like you mention, it does seem like quite a bit happened internally to cause this cascade. I'm sure much of it is technical. Once again, the modality of the product they are reaching for is brand-new and will not be a walk in the park for development and implementation. I agree that the ultimate impression of the new release "range" is that they have finally reckoned with this. But I'm also sure that much of it is also related to people. The VocaTone blog post reveals enough of that on its own.
I truly, truly hope... that this year, they'll be able to use this new corporate structure to truly do what they intend to do with this project, communicate with their investors, and stop beating around the bush. It can only benefit them to keep it real. And yes, that they loop everyone in on an "official", set form of communication.
I think we can kinda say that it became some kind of development hell type of project lol
I imagine it at 7 years of dev still not released
I imagine it at 7 years of dev still not released