I like the idea of a website with ratings, but this sounds like the opposite of quality control. It would ensure that the vast majority of voicebanks (free ones) would be lower quality.
Well, that’s the price of free content, you can always find something good, but a lot of them will be bad. If you want something really good, you can always pay for it.
The ratings is there to help you find the good things in the free tab, and the comment section helps the creators knows where to improve their VBs
Okay, but you’ve missed the point, because Kanru’s concern is with quality control and maintaining the integrity of the brand.
Yeah, I took some time to think of it and it’s really difficult to put quality control for free content even with ratings and comments, but the point of the Community Creation is that everything there is not official.
If the public have this power to create a voicebank, the only thing that Dreamtronics can put it’s sugestions of recording settings and lists, and if they create a website to put all this content all together they need to put a warning in the telling that everything that is not an official content may not have the quality what the company seeks.
If the VB creators that is not attached with some company want to put an official label in their products, Dreamtronic can create something like the Steam’s Greenlight that needs to pass through the comunity control and Dreamtronics’ quality control.
About the free VB software, I think they can be only free for whoever payd the SynthV software. They just need to put their key and done.
Well, everything here is just sugestions, hopes and dreams.
I feel I should chime in on the matter because I have several years of experience with plugins from other companies and anything that allows for custom creations, and I have been using Vocaloid products for about four years, although I don’t mean to say that to brag.
In my case, custom voicebanks are a big no for me. Open models never work well because of many reasons and I’m going to list down all of my experiences with open models.
Let’s start with Positive Grid’s Bias Amp. Just like the model you mentioned, most amp profiles can’t be accessed unless you pay for the full version, which I ended up doing, and the plugin itself just eats up a lot of processing power and for something that does what it does, it’s really overpriced and overrated. Yes, you could make your own unique amp models, and yes, you can try to match any of the amps you own, but the features the plugin has to help you match the amp is very limited, and the parts you can select to build your own custom amp are very few, making the entire feature redundant. Also, the only amp profiles you could find that are of the quality you hope for are the official endorser profiles, and there are very few of them, and often you don’t even get the amp you’re looking for. If there were more features and parts available, would things be different? Not really, because most users build the same thing. Why is that applicable with the virtual singer market? Here’s a little music industry trick that very few people know: It’s not that hard to have a similar singing voice as another singer. This is why several Vocaloid developers have been criticised for building banks that sound identical to another bank, because the VPs’ singing voices inadvertently sound the same. Even amongst Vocaloids you can’t tell some of them apart. Now imagine that but multiplied by who-knows-how-much. (And I can even bring up the story of an old prof of mine that had to deal with a videogame’s custom song feature and he got fed up with the whole system.)
Next is something I have to say about Kontakt. Yes, you could make a commercial Kontakt library, and I know another professor that does this, but he told me there’s this big NDA you have to sign and you have so many things to waive up, and because of the fact that you need a lot of knowledge in code and there’s a lot of QC that’s going on, he can’t do any of the work alone and he has to build a team, and you know what that means… Lots of $$$$ just to build a proper Kontakt library. Because most of the folks in the virtual singer community are just hobbyists and it’ll cost the same ridiculous amount of money to build any library/bank that is up to the proper standards, at best it’ll be really hard to find enough banks that’ll pass the quality check. I actually have a lot of respect for UTAU users because any and all things in UTAU are really difficult to wrap one’s head around, and the fact that they understand any of this at all is just amazing. And then comes a virtual singer platform where you have to know how to code well and sing well. Yikes.
Now let’s talk about Reaper. Yes, the infamous Reaper the DAW. Where do we even begin… Everyone knows it as the cheapest DAW in the music industry and it’s often laughed at, albeit not as much as Audacity. It claims to have a limited time trial available for people who aren’t sure if they want to buy it yet, but here’s the thing. There’s nothing stopping them from continuing to use the said trial version even after the trial period ends. Many would argue that you just have to make sure you’re not making money with the said DAW, but to tell you the truth, as long as the user keeps his/her mouth shut, it’s going to be really hard to find out he/she was making a lot of money with a trial version of a DAW or any product for that matter. There’s a reason why a lot of the newer plugin and other software developers don’t even offer up a trial/demo version at all and you can’t even use the product unless you fully buy.
Last but not least, there have been plenty of drama amongst people who have actually been part of official Vocaloid development that I’m not going to get into, but they’re all the more reason why I personally feel that making any development tool open to the public will just open doors to more controversy. I really feel Synthesizer V has a lot of potential and I don’t want this franchise to be another franchise that gets ruined by drama that goes around it.
Okay, sorry for double posting here but @Villano did bring up a few things that actually got me taking a few steps back and starting to rethink my words a little bit.
He mentioned something about ratings and I wasn’t able to think it through last night but I actually realised something.
That same prof that runs a company that makes piano Kontakt libraries that I mentioned last post? Yea, I just remembered that he also runs another company called Plugrater. What’s interesting about Plugrater is that they’re a plugin review site that not only rates well-known plugins and Kontakt libraries, but I also found out that half of the reviews they have are of plugins, libraries, and virtual instruments I’ve never heard of, and some of these are ones that these guys speak highly of.
I started thinking that… maybe… just maybe… as Villano said, we could set up a some kind of a new site where we have all of the virtual singer voicebank reviews, or maybe it could be all SynthV voicebank reviews. If that were to happen we would absolutely have to open up a custom voicebank development programme where the aspiring developers would take several courses on how to operate the engine properly and how to actually get the best out of the engine and then the voicebanks created from the said campaign would be put up on that review site for everyone to be better educated before buying. If a custom bank campaign didn’t happen and we could only review official banks that kind of defeats the point of having a site like that unless there were a lot of banks in the franchise…
I do realise that lots of voicebanks are going to sound really similar to each other, but that hasn’t stopped the Vocaloid franchise from making new voicebanks, has it… O.O Maybe some of the reviews are going to sound like that of angry gamers, but I did realise I tend to see such reviews almost anywhere. And really, if you look back, you can’t expect all controversies to be avoidable… Maybe I got too caught up in the moment and not realised some of these people might be on to something. >.<;;
(Responding to Villano’s suggestion…)
A service like Steam Greenlight might not be a good idea. When Steam Greenlight still existed, it was criticized because certain developers used tactics, like bribing people for votes, to get their games greenlit (instead of people voting honestly). Steam Greenlight wasn’t a bad idea, but it was abused so much that Valve replaced it with something else. If Dreamtonics decides to allow for custom voicebanks and uses a similar system, that is something to consider.
(Edit: fixed a sentence that didn’t make sense)
I think people who want to have a custom voicebank should contact the developer to pay to make one, instead of using the development tool, this will help to maintain the voicebank’s vocal quality and keep the reliability of the software.
(Translate with translator)
Being completely honest I don’t understand why this discussion is still going on…? SynthV is a professional software and interested parties can contact Dreamtonics to pitch their voices like with any other commercial synthesizer. It’s been made clear before that it’s unrealistic to expect this service to become free any time soon or for it to become more publicly-facing due to conditions and worries Kanru has stated before. Anyone can make a custom voicebank if they have the money to develop for it (and realistically speaking you should have a decent amount of funding if you want to create a voicebank with good quality output. If you don’t, UTAU is getting an update eventually and that’s the best you’ll get). Peer-review services don’t work from a development perspective unless all those reviewing have experience with developing voice databases at the same level or higher, otherwise it easily becomes a popularity contest and less focused on a voice database’s actual qualities.
It doesn’t have to be all people with experience with developing voicebanks though. If you look at Plugrater, very few of the staff have any experience with developing plugins or sample libraries let alone virtual instruments, but they’re all really experienced with engineering as well as using many of the said plugins, libraries, and virtual instruments. I may seem like I only have four years of experience writing original songs, but I’ve been recording and mixing my own songs since 2012 and I ended up hitting the reset button in 2015. That’s at least 7 years of writing, recording, and mixing that I have under my belt. Does that make me qualify as a reviewer? Probably not, but the point I’m making is that users who have a lot of experience with using the said products can also tell the good ones from the bad ones.
Also, I mean, if you really think from multiple angles, any “decent” amount of funding with any audio product will most definitely amount to a minimum of six figures, and I’m not even exaggerating, so by your logic it would imply that we can’t have anyone else developing voicebanks for SynthV at all, which I don’t have a problem with. I was just suggesting some ways people could go about making custom voicebanks for SynthV if they wanted to do it.
To be honest, seeing how far SynthV has come to be more of a professional software, I’ve just realized something… Making a FREE custom voicebank/database isn’t really a practical decision after all… Even though I would love to have a custom voicebank, in a business perspective however, it doesn’t seem to fit right for me… Otherwise, Kanru did allow third parties to have a voice currently… I know this may seem a sad conclusion for most people but you could always try to get hired or get in touch with Dreamtronics to have a custom voice made for you… Then, they would review it. If they do allow you have to pay them to get your voice working for SynthV. You have to know that the software uses NEURAL NETWORK to make the software figure out what audio should be played, overlayed, mixed, etc. Neural network can actually take days to weeks or even months to produce some decent quality! Well, in that case, if you want to have a voicebank of your own almost on par with SynthV (and its free), I recommend using UTAU with Kanru’s Moresampler resampler… I’ve tested it with Renri’s UTAU voicebank and Synth V voicebank (just that I have to edit the pitch bends to be almost as Synth V’s) and they both sound almost the same! I know some people would argue that Synth V is more user-friendly and you can input lyrics easily… but with time and effort, you can actually produce something similar to Synth V, if not almost human quality sounding. My thoughts maybe bad but that’s what I want to bring out :P.
I don’t think it’s not a practical decision, but if anyone were to do it for real it’s going to cost a lot of money, but any professional would be wise enough to understand that the more serious you are about anything, the more it’s going to cost you. It’s also why I have a really expensive bass and really expensive other bass equipment even though I’m primarily a guitarist, since you can’t expect to spend an average amount of money on a bass (let alone less) and make it sound good in a rock/metal mix where the tuning you’re using is so low that people advise you against it.
I don’t think discouraging people from making custom anything is a good idea, but it is helpful to caution them that if they want to do it for real it’s going to cost a lot in money, time, dedication, and all that.
I mean, think about it for a little bit. Kontakt is also marketed for professionals. But then why do they still allow almost anyone to make their own Kontakt libraries? Obviously the really good ones are going to cost a lot of money to make, but they’re still not actively stopping people from the idea. Should there be restrictions so that the ones who aren’t as dedicated are filtered out? Maybe, but if others are allowing the custom route, I don’t think it’s all that far from realistic to have anything similar in place anywhere else, even if a much smaller number of people were to be let into the system or there were to be some other restrictions because of trust issues or what have you (which is completely understandable because I tend to be the distrustful type myself).
Am I saying that your argument saying that SynthV shouldn’t have custom libraries at all is completely wrong and has no merit? Absolutely not. There are a lot of good points. Again, completely fine with SynthV not having custom libraries at all, but I just wanted to bring up the fact that other products that are marketed for professionals actually do have features that allow people to build their own libraries and plugins and such, even if it costs a lot of money.
Excuse my ignorance, but to make a custom voice bank that’s usable the input should be from a nice singing voice right? I’m ok, and willing to sing a-e-i-o-u and stuff into a mic to make wav sniplets but that would mean if I understand it, a bunch of notes of me which seems distasteful to me.
I’m not quite sure if I follow. There’s a lot more that goes into a proper voicebank than just a-e-i-o-u and there’s nothing wrong with anyone having an entire voicebank of themselves unless they themselves have a problem with it, but if you have a problem with it, don’t do it. I guarantee that at least a quarter of the Vocaloids available in the market are “vanity/signature” banks where all of the notes are modeled after the voice providers themselves such as Miriam, Piko, Nemu, Fukase, and so on. Nothing wrong with that if they’re willing to declare that they did stuff like this out in the open. You don’t see that every day so it’s good to see some people take pride in this stuff. (^ ^)
To sum it all up, typically I would actually take a side on the debate, but honestly, this is one of those cases where it’s whatever floats people’s boats. 🤷 I just like the fact that there’s a bank like Eleanor Forte that motivates me to pick up the guitar and start playing again.
I was using a-e-i-o-u as a placeholder for all the “snippets” (I don’t know the technical term for them) that make up a voice bank. Phonemes are they?
No, because phonemes aren’t consonant + vowel. A phoneme is essentially a consonant or a vowel. For example, if you look at the word “pepperoni,” I don’t know how it would be in arpabet but ph e p @r @U n i: would be how it would work in Vocaloid, or that’s what I think, but in this particular instance, ph is one phoneme, e is one phoneme, p is one phoneme, and so on. Every consonant sound makes up one phoneme and every vowel sound makes up one phoneme.
Hi, I been trying to reach you for a while. I was curious to know if you might want a male vocalist . My Vocal range is C2-F#4 so I a tenor vocalist. Here’s how my voice sounds with moresampler 083 before, the moresampler site shutdown.
I may not know much about, how synthesizer z works. But, I’d love to learn to help out however I can .
Why note try vcv, cvvc, or vccv phonemes? I personally to v-cvvc reclists for my recordings. To get better results like this this https://soundcloud.com/dionn-koitsu/1532725639sotamusara-zzaza
Do you have a example of your recordings for utau?
that would be interesting to record
till now i only have recorded CV banks xD
also i do think to have the ability to record your own VB for the program does add a lot of use out of the program heck look at UTAU XD
but i would put it behind a paywall that you have to use the dev kit for people to use it
the downside is that the voices companies make for synth v would soreta sink into the sea of voices people make for the program.
which also brings me to point that people will either sell or give people the voicebanks they created
they would be making money themselves using the devkit they bought
what you could do is that people can make their own VB and sell it through the site here and that the people that made the program get a cut from the sales or something
Well in the case of voicebanks, I’m pretty sure you’d have to record every consonant-vowel combination out in existence, especially with UTAU… but if you’re good at editing you might be able to get away with a few things…
Reminds me of that one assignment that my class had to do where we had to take a professor’s reading and turn it completely into something else that’s outright goofy and out of place, although I can’t say any examples due to profanity…