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

[Pelicanomicon]

This is mighty interesting software, Vocaloid finally getting some proper competition!

Vocaloid development seems to have stagnated with the release of V5 We've had no new English voices since 2018 and some prominent developers have switched to other engines. I do hope Eclipsed Sounds' campaign for a native English voice goes well!

[Pelicanomicon]

It's raining news!

Looks like the first commercial English voices for Synth V will be AH-Soft's Maki Tsurumaki English Standard and AI, out on June 18, 2021. These are based on a non-native performer, but results sound pretty good compared to similar Vocaloid efforts:

AI and Standard side-by-side
https://www.youtube.com/watch?v=VDufH655zNI&t=28s
Brief use of English, abt 20 seconds worth
https://www.youtube.com/watch?v=-fHYDHch21s&t=1m30s
AH-Soft download shop (pre-order with the blue "DL予約" buttons, Cleverbridge checkout in your language/currency):
https://www.ah-soft.com/product/download.html

Product names:

• Synthesizer V 弦巻マキ English AI
• Synthesizer V 弦巻マキ English
• Synthesizer V Studio Pro (the paid editor)

These banks will work with the free Synthesizer V Studio Basic, but certain features will not be accessible.

In case you need talking in your songs, AH-Soft is going to simultaneously release a Maki English TTS voicebank for the separate CeVIO AI editor.

[Spitfire31]

Well, if this squeaky falsetto sounds "pretty good", I have to wonder what Vocaloid sounds like in comparison…

Anyway, the fact that there seems to be a trickle of interest in producing Western voices for Synthesizer V is encouraging, at least. Hopefully, there will eventually be something useful, with some maturity and power.

[Pelicanomicon]

Eclipsed Sounds' crowdfunding for the first commercial native English Synthesizer V voicebank is now live on IndieGoGo
https://igg.me/at/synth-v-solaris

Base goal is $40,000 in 45 days. Seems doable.

Well, if this squeaky falsetto sounds "pretty good", I have to wonder what Vocaloid sounds like in comparison…

May I offer you an AI-trained Vocaloid of a famous Japanese singer who passed away 30 years ago, singing a song she never sang when she was alive to an audience of loyal fans, her friends and fellow musicians, and her own son?
https://www.youtube.com/watch?v=kgk_-1xYVz4&t=39m30s
Anyway, I should have been more clear "Pretty good" pronunciation compared to previous non-native English Vocaloids. The character of a voicebank can be changed quite a bit by messing around in Synth V's Tension and Gender automation lanes. Won't deny that the vocalsynth fandom L O V E S falsetto, tho!

[pLasterbrain]

Still waiting for a full Eleanor Forte release...

[Pelicanomicon]

Still waiting for a full Eleanor Forte release...

Have some news about Eleanor*! A few days ago, Eleanor's company Volor tweeted:

https://twitter.com/OfficialVolor/statu ... 6488153089

TLDR: Eleanor's voice provider had been flying to Japan for Standard recording sessions with Synthesizer V's programmers. Then...pandemic. Volor can remotely create an AI VB, but it won't sound the same. They want to know your feelings about it!

*For Synth V newcomers: Eleanor Forte is currently the lone English voice for Synthesizer V. It's free, but unfinished and can't be used for commercial works.

[pLasterbrain]

Still waiting for a full Eleanor Forte release...

Have some news about Eleanor*! A few days ago, Eleanor's company Volor tweeted:

https://twitter.com/OfficialVolor/statu ... 6488153089

TLDR: Eleanor's voice provider had been flying to Japan for Standard recording sessions with Synthesizer V's programmers. Then...pandemic. Volor can remotely create an AI VB, but it won't sound the same. They want to know your feelings about it!

*For Synth V newcomers: Eleanor Forte is currently the lone English voice for Synthesizer V. It's free, but unfinished and can't be used for commercial works.

Arghh... I think she sounds great as is (my other library I own rn is English Miku, so.), it's just the non-commercial thing is killing me. But news is better than no news since the last I saw people were suspecting Kanru flew the coop ^^

[Pelicanomicon]

Eclipsed Sounds has raised 33% of its goal to create the SOLARIS English voicebank!
https://igg.me/at/synth-v-solaris

They tweeted a duet of SOLARIS and the project's voice provider, Emma Rowley
https://twitter.com/eclipsedsounds/stat ... 2058663941

Again, this is a very limited test VB. The goal is to AI-train SOLARIS on enough of Rowley's data so that it sounds like her.

[Pelicanomicon]

One week later, Eclipsed Sounds has crowdfunded $15,000 of its $40,000 initial goal to create a native English voicebank for Synthesizer V
https://igg.me/at/synth-v-solaris

The team streamed a live Q&A today:
https://www.youtube.com/watch?v=ET6Kdw15L_8

imo their laser focus on a natural-sounding, high-quality single-language voice is reassuring. I timestamped questions that might be of interest to KVR users. My comments are in italics.

• 4:32 Why do the [crowdfunding] tiers cost what they do?
  (The "Tsuina-chan" mentioned is a Jpn Synth V character that had an incredibly successful crowdfunding campaign. It raised ¥10,000,000 for Standard & AI VBs in 22 days, ended at ¥15,000,000.)
• 5:25 What does the 'flexible funding' text on the IndieGoGo page mean? Will you still make her if you don't hit the goal?
• 9:30 Will you consider a Standard voicebank if you're unable to fund the AI voicebank, if it's cheaper?
• 10:50 Regarding the test voicebank...will there be any other full demo songs or is the test voicebank really only capable of singing "Miles"?
• 11:58 Do you already know what kind of features Gen 3* Synthesizer V AI will bring to the table and how SOLARIS will be able to utilize them?
• 13:03 Are you looking to show off particular parts of her voice in future demos and, in that case, do you have a certain set of genres that you're looking for?
• 14:10 How many pitches is she planned to have?
  (Several Standard Synth V VBs advertise which pitch layers they were recorded at.)
• 15:10 Does that mean minimal tuning for the AI or just realistic base vocals, how does that work? (Team talks about how Synth V Studio Pro and Basic will handle this.)
• 17:00 Do I have to buy Synthesizer V to use SOLARIS?
• 17:38 Will a Standard bank be released as well as AI or is it just AI that's planned?
• 18:25 How many songs are required to make a high-quality database for English?
• 19:18 Will SOLARIS have a free version after the product is released?
• 20:04 Can you give us any details on the license to use SOLARIS' voicebank for commercial use?
• 21:22 Are you planning on giving SOLARIS a male counterpart one day in the future, like [VOCALOID] Tonio has Prima and Sweet Ann has Big Al?
• 22:48 Can you explain the journey to make SOLARIS so far? How does one make a vocal synth?
• 25:37 Will SOLARIS come with exVoice samples?
  (exVoice samples are short WAVs spoken by a voice provider--"Hey!" "1, 2, 3, 4!" "Mic check!", etc.)
  (BTW, you can pay the crowdfund for a personal exVoice recorded by Rowley )
• 26:35 How did Eclipsed Sounds decide who was going to voice SOLARIS? What were the qualifications? And were there other singers we reached out to?
• 27:26 How does Emma feel about putting her voice onto a synth?
• 31:19 I know Emma Rowley is a very talented belter. Will that translate into Synth V?
• 32:43 What got each of you interested in doing synth voices?
• 38:00 If the Indiegogo doesn't reach its goal, do you have a limit on how much you'd be able to [raise] yourselves?
• 44:25 If SOLARIS does really well, would you consider making a Japanese voicebank for SOLARIS in the future?
• 46:12 In the event that Synthesizer V AI is upgraded to a new generation after SOLARIS is released, will SOLARIS upgrade as Saki AI has done?
• 47:12 Is there a chance you guys can do a giveaway or contest of some sort for Synth V Studio?
  (This is what AHS is doing right now, giving away box/download bundles of Pro + all 4 of Maki Tsurumaki's VBs on Twitter. Foreigners can enter.)
• 48:49 What genres would you say SOLARIS would be best for?
• 49:38 Could SOLARIS be available in a starter pack bundled with Synth V Studio Pro?
• 54:17 Does Synth V AI have good vocal fry? (They reveal Eclipsed Sounds is working on custom Synth V Studio Pro plugins for SOLARIS.)
• 57:05 Is it possible for people to submit demo songs for SOLARIS? (Eclipsed Sounds states their position on compensation for artists.)
  (It's )
• 59:28 How will the market price of the voice library be determined?
  (Japan prices run ¥8580~9680, Taiwan is NT$1980~3200.)
• 1:04:40 Will she have breath samples?
• 1:05:01 Will SOLARIS have extra phonemes?
• 1:07:48 Do you have an estimate of what her vocal range will be? Will she have a falsetto pitch or just standard? Will she have autotuning features? Will her breath samples be WAV files or programmed into the voicebank? Will she be compatible with [Synth V Studio Pro feature] Tone Shift?
  (Tone Shift is an AI parameter that alters the timbre associated with the pitch of a sung note.)
• 1:09:26 How news outlets/social media managers can get in touch with them.
• 1:14:01 How hard is it to program SOLARIS, or Synth V in general?
• 1:14:15 With the intonation of her current bank, do you think she'd be good for other languages?
• 1:35:08 Would you be open to doing what AHS wants to do and hire native speakers of other languages to make different language banks for SOLARIS?

*The question at 11:58 refers to Gen 3 AI revealed on AHS' livestream on May 13, 2021. At each timestamp, the first bit is previous gen followed by new AI.

• Ex.1: 1:07:17 (Saki AI)
• Ex.2: 1:07:55 (Saki AI)
• Ex 3: 1:09:34 (Maki Tsurumaki Jpn)
• Ex 4: 1:10:10 (Maki Tsurumaki Eng)

https://www.youtube.com/watch?v=uJTeCfL8nX4&t=4037s

[Pelicanomicon]

With 4 weeks to go Eclipsed Sounds has raised a little over 40% of their goal for the SOLARIS native English voicebank. Music Radar gave them a mention too
https://igg.me/at/synth-v-solaris

Eclipsed Sounds posted a short interview with SOLARIS' voice Emma Rowley on their Youtube channel. She seems pretty excited about it all Supposedly one of the big roadblocks for vocal synth developers is finding professionals okay with their voices being cloned/used in ways they can't control.

[Pelicanomicon]

AH-Soft has uploaded free Lite versions of Tsurumaki Maki English Standard and English AI a whole 2 weeks before the full commercial releases!

Go to
https://www.ah-soft.com/trial/synth-v.html
and check the box next to
プログラム使用許諾に同意する (the Usage Agreement)
at page bottom to access download links

The products are

• Synthesizer V Studio Basic (free editor)
• Synthesizer V 弦巻マキ English ライト版
• Synthesizer V 弦巻マキ English AI ライト版

You'll have to update Synthesizer V to 1.2.4 to install

Lite voicebank restrictions apply: works must be not for profit and if published you must note "This work uses Synthesizer V Tsurumaki Maki English Lite" or "This work uses Synthesizer V Tsurumaki Maki English AI Lite"

Quick impressions

Standard is tbh...uncanny. Maki Std Lite is way more robotic than Eleanor Lite. Note transitions are abrupt and the pronunciation is like Crypton English Vocaloids. Maybe full Standard will be better cause it'll have the missing pitch layers?

On the other hand, AI! AI Lite is really promising! More musical, as opposed to Cyber Diva's talk-singiness. Enunciation feels more natural than Eleanor, probably cause of the training on whole songs.

I did find a consistent bug in the way Maki AI pronounces single-syllable consonant+[r]+vowel words like "grow", "fright"--the initial cluster becomes unnaturally prolonged. But you can work around it by splitting the note--"f r / ay t" or "f r ay t / -" and tweaking phoneme sliders. The bug doesn't happen if it's a two-syllable word ("grindhouse").

Pronunciation issues aside, Standard has an edgy sound that might be good for rock?

[Pelicanomicon]

Synthesizer V Studio updated to 1.2.6, includes a fix for crashes upon choosing an export folder

[Pelicanomicon]

The SOLARIS campaign has crowdfunded 45% of their native English voice but must now raise over $1000/day til their deadline of June 28

[Spitfire31]

What grown ups can take this 'Solaris' campaign seriously?

Demos will be posted on Tik Tok! And 'awards' consist of garbage cartoon trinkets. Looks like a product for preteens that can trick 125 bucks from their parents for the digital version that may be available in Feb 2022, if…

It looks and sounds like a scam. The only thing that rings true is the name of the gang behind it: "Eclipsed Sounds". The kids don't seem to realise that 'Eclipsed' means 'obscured', 'in the shadow of', 'deprived of significance'. Maybe Eclipse Sounds was taken?

Caveat emptor.

[Pelicanomicon]

It looks and sounds like a scam.

The Synthesizer V devs have endorsed the campaign tho
https://twitter.com/dreamtonics_en/stat ... 9092802562