I am in the process of completing an album. Apart from my vocals on the first track, all other tracks use vocals from Kevin, Weina, Solaria, and Asterian. I recognize that the license permits their use on commercial projects, but I’m curious about how others are attributing the vocals. I’m reticent to say that they are AI voices because they do require quite a bit of programming and are modeled from human voices. Some possible options:
1- Don’t even mention who is singing. This seems problematic since I don’t sound like any of them and the gender doesn’t always match.
2- Say that the other vocalists were “brought in” to perform the songs, but don’t mention their names.
3- Make up something like "Members of the AI family- Wendy AI, Solaria AI, Asterian AI, and Kevin AI.
4- I’d prefer to reference the actual human singer- Like Weina Hu, Eric Hollaway, etc but I believe that’s prohibited.
5- Simply say the vocals were created on a vocal synthesizer by Dreamtonics.
The issue is I’d like the focus to be on the music more than the technology. I’m curious what others are doing. Amazing product. I’ve been creating music for 45 years, but because my voice isn’t that strong, I’ve been unwilling to create a professional project. This software has helped to fulfill a lifelong dream. Thanks.
You can refer to these pages for the terms of use of Dreamtonics and Eclipsed Sounds voices respectively:
If you do choose to attribute the vocals in some way, make sure you take note of this clause (present in both license agreements with slight wording differences):
3.2 Under the condition that:
(1) When Presenting a Rendition, the User must not use names other than designated by Dreamtonics to identify the SVD.
This means on streaming services, you would include the product name in the metadata as a vocalist. “Solaria”, “Kevin”, etc., but you strictly cannot credit the original human voice provider because that could qualify as impersonation.
On other platforms like YouTube where you have a text description, you can say any variation of these examples (or come up with something similar):
“Vocals created with Synthesizer V”
“Vocals created with Solaria and Asterian for Synthesizer V”
“Vocal: Solaria, Kevin, Weina”
Personally I find that describing the product as an “AI vocalist” does a poor job of conveying what it does in practice, and is likely to result in people making incorrect assumptions. The software works functionally the same way as pre-AI vocal synths did, just using newer tech to produce higher quality output.
A lot of people are at least vaguely aware of sample-based instrumental synths – or at least the idea that you can download a virtual piano on your phone or tablet and it makes piano sounds without needing a physical piano in the room – so I think describing it as a “vocal synthesizer” tends to help in conveying the idea more accurately and gives people a more helpful starting frame of reference.
Can anyone explain what this paragraph is trying to convey?
“For commercial works using the Synthesizer V SOLARIA character mascot and not the synthesized voice, derivative products may be sold without permission or profit limit by individual persons or non-business entity production groups. Users are not required to use the character design in any way.”
The product has two components to it:
- The synthesized voice that the software produces
- The character used to market the product (sometimes this also includes the name of the character mascot, but that distinction is not particularly relevant here)
If you make a song with Solaria, you’re using #1, if you make a keychain of the Solaria character mascot that’s #2, and if you release an album and put the Solaria character mascot in the album artwork, that’s both.
That clause refers to cases where someone is making merchandise featuring Solaria (whether using the Eclipsed Sounds character design, or a new design for the same character), but where the sales are not related to a music release.
In this case Eclipsed Sounds is saying “hey, feel free to use our character design”, because other companies have much more strict restrictions on the character, and it is common to be allowed to use the synthesized output without permission while needing permission to use the character.
Totally agree with the quote below from Claire. If one starts crediting anything as “AI” people tend to conclude it is completely artificially generated - that you did not even write or perform the song/parts. As the creator of the music, I want to avoid that misunderstanding.
Two different examples I’ve used - (still deciding the best tact)
“Vocals performed using Dreamtonics Synth V/Natalie”
“Vocals - Natalie (Dreamtonics)”
Both are OK I think, legally/EULA speaking. #1 is very specific. #2 is a bit more vague. Someone that doesn’t know what Dreamtonics is might assume it’s a vocal group or singer representation/company. This could be good to de-emphasize the focus on the tools and spotlight the music itself (my personal preference) - with a small risk of someone feeling duped into thinking it’s a human. Splitting hairs - but that would be their assumption and no fault of yours. 
YMMV.
Personally I find that describing the product as an “AI vocalist” does a poor job of conveying what it does in practice, and is likely to result in people making incorrect assumptions. The software works functionally the same way as pre-AI vocal synths did, just using newer tech to produce higher quality output.
A lot of people are at least vaguely aware of sample-based instrumental synths – or at least the idea that you can download a virtual piano on your phone or tablet and it makes piano sounds without needing a physical piano in the room – so I think describing it as a “vocal synthesizer” tends to help in conveying the idea more accurately, or at least gives people a more helpful starting frame of reference.
That’s exactly, what you will notice on many commercial releases on Spotify & Co.
PS: And that’s what I will do.
Thanks for all the great suggestions. I anticipate I will simply define the vocalists as Solaria, Weina, Kevin, Asterian. Since I only sing on one track, I’m wondering about not even using my name as the primary artist but as part of a project (similar to The Alan Parson’s Project). I will definitely stop referring to Synth V as AI. While I know I can’t list the actual human vocalists behind Synth V in the credits, if people ask how the product works, I assume it would be ok to mention that it was created by modeling actual human voices.
I bought Synthesizer V just recently, and was aware of the terms mentioned above. They are perfectly OK with me.
However, this type of policy may work both ways, of which I am quite sure that Dreamtonics and their associates are perfectly aware:
By comparison, traditional sample libraries can be used for commercial purpose without the need of referring to them in any way whatsoever during the closing credits, being it Vienna SL, EastWest, Cinematic Studios, CineSamples + + +.
I think the reason for this is obvious; if a company would choose a solution like the one we discuss here, composers dealing with cinema, TV-productions or major CD-releases would not use it.
There has hardly ever been any reference to a sampled orchestra or sampled singer during any Closing credits, as far as I can recall. Nor do I think there will be any in the near future.
Still the vast majority of film and TV-music is made via Virtual Studio Technology.
On the other hand, to my knowlege Synthesizer V has no real competitor at the moment. I have been using traditional sample libraries for decades, from symphonic and choral to rock, but Synthesizer V is a giant leap forward. It’s a breakthrough.
This thread is about someone asking how to attribute the vocals in a context where they specifically want to provide information to their audience about the vocals.
For most voice database there is no requirement to provide any attribution whatsoever. The terms don’t require you to provide any credit, they just prevent people from lying about how the vocals were created.
There are some third-party voices that do have more strict terms of use, but these are primarily ones targeted at the Japanese market where there is 15+ years of precedent.
[quote=“claire, post:2, topic:11834”]
This means on streaming services, you would include the product name in the metadata as a vocalist. “Solaria”, “Kevin”, etc., but you strictly cannot credit the original human voice provider because that could qualify as impersonation.
On other platforms like YouTube where you have a text description, you can say any variation of these examples (or come up with something similar):
This means on streaming services, you would include the product name in the metadata as a vocalist. “Solaria”, “Kevin”, etc., but you strictly cannot credit the original human voice provider because that could qualify as impersonation.
On other platforms like YouTube where you have a text description, you can say any variation of these examples (or come up with something similar):
“Vocals created with Synthesizer V”
“Vocals created with Solaria and Asterian for Synthesizer V”
“Vocal: Solaria, Kevin, Weina”
Thanks a lot for your reply. To me, this was a Clarification; so there is no need to refer to any of the examples you listed; one can choose to leave the performer out if needed, but not make up a fake performer.
Why? That’s not needed, except when using the free version, but that is not allowed for commercial releases, which every streaming service release (e.g. Spotify) would be.
I also would never give credit to an unknown studio musician, only to famous artists, that might be involved, which usually is marked with a “featuring” remark.
To be clear, the Basic edition can be used for commercial work, it is the voice database alone that determines that aspect of the usage terms.
Vocal synths are not just instruments; they’re also marketing tools. Nobody cares which violin VST you use, but how the vocals sound is crucial to whether or not someone listens to a song.
Many people specifically search for music featuring Solaria, for example. Adding the Solaria Spotify profile or stating “ft. Solaria” in a YouTube title will give your music more discoverability.
Ok, understandable, but I myself wouldn’t want that and as it seems, I’m not alone 
On Spotify - Solaria linked above is listed as a “Verified Artist” - which is 100% false information.
It seems there are those jumping on that bandwagon (guessing it’s in hope of more clicks) - but a synth sound tool is not an artist, person or creator of anything. Misleading at best - fraudulent at worst. And if the posted work isn’t liked/interesting to the listener, it makes “the artist” Solaria something to actively avoid playing other works mentioning the voice DB, which could actually hurt other user’s chances at clicks.
Where it does make sense to feature a Voice DB name, is in the communities of voice singing synthesis - but also outside those areas publicly, when showcasing the possibilities. Not as an artist, but as a tool.
YMMV
Correct and I absolutely hate Spotify for two reasons alone:
- still no HD content, only crappy MP3 quality
- it’s a complete Kindergarden
Spotify does not offer any means to create a public profile for anything other than an “Artist”.
The reason it’s verified is because the profile was created by Eclipsed Sounds at the request of users. Having an official profile to link to is much better than having a dozen unofficial ones (if you do a couple searches, you’ll see plenty of user-made artist profiles for vocal synths that were not created by the actual company, so “there shouldn’t be a profile at all” isn’t really an option in the current Spotify ecosystem).
Again, linking your song to it is wholly optional. Nobody said you had to do it (again, barring the exceptions mentioned above where the products are directed to a market where vocal synths have been commercially viable for over 15 years and there is plenty of precedent for how to handle the topic of attribution, however this thread is specifically about Kevin, Weina, Solaria, and Asterian).
Nice topic James - I am in a similar position - looking to release some songs through either CD-Baby - Tunecore etc, so I was wondering what you finally did? I do sing; even if it’s only backing vocals but wondering what to put for the artist? I believe with CD-Baby if you did put say Solaria - then there would be a search for that name on playlists - databases etc - so in my eyes that might confuse streaming royalties etc - appreciate any feedback / comments on this…Thanks
I ended up simply writing “all other vocals created using a vocal synthesizer.” If someone asks, I’ll give them more details. As to releasing it, I’ve been researching release strategies and have only heard bad things about Tunecore. Your mileage may vary though. It
Thanks for that - (looks like your reply got cut short?) yes I am currently researching online streaming distributers and I also have crossed TuneCore off the list - can I ask who you have signed up with and do you have any advice about them etc?
I’m still working on some details for the release, so I haven’t chosen a distributor yet. Looking into details related to royalties was far more complex than I realized, so I’ll save you some time there. This is what I figured out on that:
There are 4 royalty streams- 2 for the composition and 2 for the sound recording. In order to receive all your royalties you need to sign up with (1) a PRO- like BMI or ASCAP in the US, (2) a mechanical royalty collection agency, like themlc.com (3) a distributor will collect your master recording royalties (4) Soundexchange to get your digital performance royalties.
What distributor you choose will depend on what works best for you. If you want to release a lot of songs over the next year, then distrokid might be best for you, but you have to keep paying each year and the extras will cost more. I’ll probably go with distrokid for the first year, since I’m doing a staggered release, and then go to CDBaby after that.