Synthesizer V - Singing voice synth for Win/Linux/Mac (Free & Payware)

[Pelicanomicon]

Will the cross-lingual technique allow, for instance, French, German, Danish, and so on versions?

Synthesizer V officially supports Chinese, Japanese and English synthesis, languages the chief developer is fluent in. I have wondered for a while if that solely determines what the synth will sing in.

BUT, if you're feeling cussed enough you can absolutely make vocal synths do things they were not designed for.

Quadimension Chiyu Standard singing in Russian
https://www.bilibili.com/video/BV1pe411x7k6

Saki AI singing in English before Cross-Lingual Synthesis was implemented

Qingsu AI singing in Polish. Going by the phonemes in the piano roll, this uses Chinese language synthesis.

btw if you're interested in wider language support, another program to keep an eye on is Maghni AI from Misbah Studios / VocaTone Studio. They are specifically focusing on multilingual voices--German, Icelandic, Arabic, Swahili, Esperanto...everything. It is in extremely early development.

I just found this thread despite it being around for years it seems. This product is rather hidden in plain sight. And it is a bit difficult for me to navigate the Japanese language website to understand what to download to get this work.

Yea that's one of the annoying things about the official Dreamtonics site, at least on my browser. It always seems to default to Japanese display and the language toggle doesn't stick.

Here is the English product page if you're new to Synthesizer V Studio
https://dreamtonics.com/en/synthesizerv/

Here is the freeware downloads page
https://resource.dreamtonics.com/download/

Download Eleanor Forte AI Lite voice from AniCute
https://www.anicute.com/

[ZincT]

Thanks for the further information and updates @Pelicanomicon

The more I hear of Solaria the more I want it! The only webpage I can find is here

https://www.eclipsedsounds.com/solaria

Do you know if it will be on pre-order there or through resellers such as Anicute?

[Pelicanomicon]

Do you know if it will be on pre-order there or through resellers such as Anicute?

Hello, @ZincT! Eclipsed Sounds tweeted that the order page will appear on their official website on January 1, so yes it will be through them. During a livestream when the crowdfund was going on they mentioned exploring sales partners, but I guess nothing is in place for now.
https://twitter.com/eclipsedsounds/stat ... 8601128960
Eclipsed Sounds website: https://www.eclipsedsounds.com/

Here's an end-of-the-year present from Quadimension, an extended demo of Cangqiong English and Muxin English singing "Lonely Warrior" (孤勇者). Muxin is one of only two male voicebanks currently sold for Synthesizer V; the other is VOLOR/VirVox Genbu Standard.
https://www.bilibili.com/video/BV1V34y1z7Mt

Per Quadimension's BiliBili blog, if testing goes well they hope to release the updates like... *checks DeepL translator* *doublechecks with Google Translate* ...uh, "like a flood". Yup, let's go with that. HAPPY NEW YEAR, EVERYONE

[ZincT]

Thanks @Pelicanomicon and a very Happy New Year to you also!

I was looking forward to hearing the end of the earlier Cangqiong demo and it is awesome!

[Spitfire31]

Happy New Year, Pelicanomicon, and thank you very much for acting as our eyes and ears in this rapidly evolving field!

The 'Lonely Warrior' demo is impressive, especially in highlighting the dynamic abilities of these AI voices. Good times ahead in 2022, at least in this particular domain!

[Scotty]

Does anyone know the specifics regarding licensing and authorization. I am wondering if we can have more than one install to support desktop and laptop at the same time.

[VitaminD]

What is the licensing for including vocals produced by Synthesizer V in commercial releases?

[Pelicanomicon]

Does anyone know the specifics regarding licensing and authorization. I am wondering if we can have more than one install to support desktop and laptop at the same time.

Sadly the license txt in Synthesizer V's install folder says no:

8 The license granted by us to the user allows the product to be installed only on one computer. The user shall not install the product on multiple computers with a single license.

9 You may transfer the product to another computer without obtaining a new license. However, you must delete the product completely from any computer on which it is already installed.

[briefcasemanx]

What is the licensing for including vocals produced by Synthesizer V in commercial releases?

lol it better be "you paid money for it so use it on whatever the hell you like"

[Pelicanomicon]

What is the licensing for including vocals produced by Synthesizer V in commercial releases?

Licensing depends on the voicebank used. Each vb's install folder comes with a license.txt. For example, Dreamtonics' Saki AI's license is the same as the Synthesizer V license txt:

10. Except as otherwise specified, you may use the synthesized sound generated by using this software, whether commercial or non-commercial, free of charge.

11. Except as otherwise specified, the user may indicate that Synthesizer V Studio(or the voice database for Synthesizer V Studio) has been used in your productions.

---

Generally it's the use of the vocal character's name and accompanying avatar/likeness that requires you to immediately contact the developer. Ex: AHS' Maki Tsurumaki's voice license txt is the same as above, and when you visit their website:
https://www.ah-soft.com/synth-v/maki/english.html#faq

「Synthesizer V 弦巻マキ English AI コンプリート」で作った作品を商用利用したいのですが？
「Synthesizer V 弦巻マキ English AI コンプリート」を使用して生成した合成音声は、商用または非商用を問わず、無償で利用することができます。
Synthesizer V 使用許諾と、キャラクター使用ガイドラインをよくお読み頂き、この内容を順守してご利用ください。
I want to use my work created with "Synthesizer V Maki Tsurumaki English AI Complete" for commercial purposes.
The synthesized voice generated by "Synthesizer V Maki Tsurumaki English AI Complete" can be used for commercial or non-commercial purposes free of charge.
Please read the Synthesizer V License Agreement and Character Usage Guidelines carefully, and use them in compliance with these terms.

And if you go to the Character Usage page
https://www.ah-soft.com/licensee/

VOCALOID、Synthesizer V 製品全般
製品のキャラクター(絵・名称)を使用せず、製品で作曲した楽曲のみを配布される場合は使用許諾契約書に則っていただいている限り、営利非営利問わずご連絡頂くことなくご利用頂けます。
VOCALOID, Synthesizer V products in general
If you want to distribute only the music composed by the product without using the product's character (picture or name), you can use it without contacting us, regardless of commercial or non-commercial use, as long as you follow the license agreement.

AHS publishes voicebanks for a lot of other companies, so those characters (Koharu Rikka, Kyomachi Seika, etc.) may have their own restrictions.

---

Previously I applauded Audiologie's transparency re use of their voices, but after reading the license txt there is a discrepancy

https://audiologie.us/pages/about-us

How is commercial use of voice databases and/or character designs for AUDIOLOGIE products handled?
All voice databases sold by AUDIOLOGIE are free for commercial usage, but commercial use of character image likeness requires permission from AUDIOLOGIE. For commercial usage of character designs, please email info[at]audiologie.us.

ANRI license.txt

2. Natural person users can use the synthesized voice generated by or role copyright image included in this product under non-commercial or commercial situations for activities, creation, work spreading, publication, etc.
3. If the natural person users apply the synthesized voice generated by or role copyright image included in this product for commercial purpose, and the situation involves more than USD 5,000, Or if the user is Corporation or legal entity, the obligation to notify AUDIOLOGIE will occur. AUDIOLOGIE will sign additional User License Agreement with such users according to the users’ use purpose (licensing fees might be produced according to the agreement situations).

I sure hope Eclipsed Sounds' license for SOLARIA is less restrictive than this.

---

Animen/VOLOR Eleanor Forte AI's license.txt

2. Natural person users can use the synthesized voice generated by this product or Copyrighted Character Image included in this product under non-commercial situations involving creation, distribution and publication of works and participation in events with no obligation to inform AnimenJP in advance.
3. Natural person users can use the synthesized voice generated by this product or Copyrighted Character Image included in this product under non-commercial situations and commercial situations involving no more than USD 1000 of net revenue per year for uses involving creation, distribution and publication of works and participation in events with no obligation to inform AnimenJP in advance.

This official Synthesizer V forum thread gives the contact information for Animen as volor[at]animen.com.tw
https://forum.synthesizerv.com/t/transp ... terms/1105

Anicute/Animen business page
https://www.anicute.com/pages/faq_help

---

Translations done with https://www.deepl.com/translator

[markmann]

Thanks Pelicanomicon for that helpful licensing information. I just finished my first song with Eleanor Forte and the people I've played it for don't hear anything obviously artificial in her voice. I don't tell them it's computer generated before they listen. And I haven't even started to tweak the various parameters yet. This is just her default articulation with a reduced and delayed vibrato.

It's eerie how evocative she can be. She places emphasis correctly on multi-syllable words over 2-3 notes and distributes complex one syllable words naturally over single extended notes (the word "thirst" for example). She also handles contractions well ("they'll" for example). She may not sound 100% "natural" in all situations, but her peculiarities almost sound like an accent which can actually be interesting. Generally she just sounds right. Amazing.

I did notice one issue in Studio One 4.6.2 (Mac). Synthesizer V doesn't seem to report the buffer reliably for compensation. So if you align the vocal notes and then change the buffer size, the timing will be off. The workaround is to render to an audio track once you get a vocal part aligned correctly with a certain buffer setting. Has anyone else on other DAWs experienced this?

[VitaminD]

Previously I applauded Audiologie's transparency re use of their voices, but after reading the license txt there is a discrepancy

https://audiologie.us/pages/about-us

How is commercial use of voice databases and/or character designs for AUDIOLOGIE products handled?
All voice databases sold by AUDIOLOGIE are free for commercial usage, but commercial use of character image likeness requires permission from AUDIOLOGIE. For commercial usage of character designs, please email info[at]audiologie.us.

ANRI license.txt

2. Natural person users can use the synthesized voice generated by or role copyright image included in this product under non-commercial or commercial situations for activities, creation, work spreading, publication, etc.
3. If the natural person users apply the synthesized voice generated by or role copyright image included in this product for commercial purpose, and the situation involves more than USD 5,000, Or if the user is Corporation or legal entity, the obligation to notify AUDIOLOGIE will occur. AUDIOLOGIE will sign additional User License Agreement with such users according to the users’ use purpose (licensing fees might be produced according to the agreement situations).

I sure hope Eclipsed Sounds' license for SOLARIA is less restrictive than this.

---

Animen/VOLOR Eleanor Forte AI's license.txt

2. Natural person users can use the synthesized voice generated by this product or Copyrighted Character Image included in this product under non-commercial situations involving creation, distribution and publication of works and participation in events with no obligation to inform AnimenJP in advance.
3. Natural person users can use the synthesized voice generated by this product or Copyrighted Character Image included in this product under non-commercial situations and commercial situations involving no more than USD 1000 of net revenue per year for uses involving creation, distribution and publication of works and participation in events with no obligation to inform AnimenJP in advance.

This official Synthesizer V forum thread gives the contact information for Animen as volor[at]animen.com.tw
https://forum.synthesizerv.com/t/transp ... terms/1105

Anicute/Animen business page
https://www.anicute.com/pages/faq_help

---

Translations done with https://www.deepl.com/translator

I appreciate the detailed response back. And this is what I was concerned about. 1000 USD per year is tiny for a commercial release, for example with the Eleanor license for vocal use. It makes me wonder what the terms are for large commercial releases? As you say they want you to contact them and handle it on a case-by-case basis. (seemingly to ensure they get paid a percentage of sales of your release should it hit it big)

Could be cheaper to just hire a real person and form a license agreement with them? Interesting. Synth V still could be useful for my purposes in coming up with vocal ideas though as well as creating guides for vocalists maybe.

[MJACau]

Some of the dumbest legalese I have seen, An analogy would be like me buying Serum and if my song is a hit, Xferrecords can charge me $1000???! Get lost!
It's bad enough with the one activation limit, and say my computer's hard drive dies there goes that activation and my money and license.

In saying that the software is good for what it does, the licencing is anti-consumer.
It makes Emvoiceapp more compelling licence wise, even though it's online.

[Spitfire31]

All this licensing stuff sounds like a theoretical exercise to me. How can anybody prove that it's Eleanor Forte singing and not my girlfriend?

With the amazing development we're seeing, er… hearing, it will soon be more or less impossible to tell a Homo Sapiens singer from a digital model, if it's well programmed.

Personally, I was gobsmacked by Cangqiong, a Chinese voice singing in English, thanks to Synthesixer V AI v.1.5 with cross-lingual capacity:
https://twitter.com/Creuzer1/status/1474262759676149767

I think that Emvoice has some serious catching up to do…

[zerocrossing]

i know, for myself, i would never use any plugin, no matter how good, to replace a real singer. i want it to be another effect, an 'android' voice, or... some combination of those things.

i can see, tho, for song demos, the benefit of having a 'better' voice...

I was working on a track once and I wanted a female vocalist on it. I put out some feelers and one woman who was the mother of one of my daughter’s friends chimed in and said she’d love to do it. She tells me that she’s a professional singer. I asked her about pay and she said she didn’t want pay, she’d do it for fun. I told her that it was to be part of a commercial soundtrack and there could be money made indirectly from the recording. She said not to worry.

First sign of an issue was that it was nearly impossible to get her to commit to a time. I sent her a recording with scratch lyrics and told her to let me know if the key was fine. Silence. A few pings later and she said the key was fine, but she’d let me know in a few more days when she could come over. A few weeks passed and she finally committed to a day.

She shows up and it’s clear to me that she has never listened to the track or read the lyrics. She didn’t have a lot of time and most of it was spent learning the song and figuring out what key to transpose it to. She’s also allergic to my cat, so a lot of time is wasted drinking tea. I gave her a performance release to sign and she said “I’ll read it tonight and sign it.” I assured her that it was just a boilerplate form that relinquished rights to the recording of her performance, but she said she wanted to read it.

Weeks go by. I’m now a full month past the time I wanted to get the project done. It’s mixed and ready to go. Finally I offer to take her out to lunch so I can pick up the release form. She shows up, eats… and doesn’t have the form. I knew something was up so I brought another copy and she tells me she hasn’t read it yet. (BTW, she’s literally a stay at home mom who’s single kid is in school from 8 to 3 every day.) She promises to read and sign it later that day and bring it to me.

I never see her again. At that point I gave up on her and found another friend who helped me out. Not quite as good a vocalist, but fine. I had to change the key yet again, which ate up a lot of time I should have been spending doing basic polish.

The moral of the story… unless you’re super tight with someone, it’s probably best to use technology because people can be flaky as fu<k.