SynthV Synthesizer V Discussion Thread

[Scarlet Illusion]

Now that I have the full editor and a full voicebank, I’ve become very interested in all things SynthV, haha. I’d be curious to see what they do with AI.
Perhaps they only announced it in Chinese because that will be their AI focal point for now? Or perhaps they’re going to play the waiting game and see how it goes over with the Chinese market and then announce it later in other languages? Just speculating here.

 

[Silver1063]

The post seems to have been removed now that people are finding it... intereeesting.

 

[xuu]

Yeah. I think it might not have been... intended to be posted. I'm aware of it a bit, but I'm wondering what exactly will be announced.

 

[Prism]

Wasn't he working on a synth based on Kiritan database I wonder if it has anything to do with that

 

[Stardust]

I would not mind if they had AI versions of VBs, perhaps bundled with the sample-based VBs or vice-versa.
I would really prefer they do not lock any VBs behind the AI version personally.

 

[uncreepy]

I hope this update brings growl back.

 

[mahalisyarifuddin]

New Generation AI Singing Voice Synthesis Software | Dreamtonics株式会社

Press Release October 30, 2020 Dreamtonics Co., Ltd. New Generation AI Singing Voice Synthesis Software Synthesizer V AI Expected Arrival Early 2021 Tokyo-based Dreamtonics (representative: Kanru Hua)

dreamtonics.com

 

[Rylitah]

Demo linked in the Dreamtonics Blog:


More examples and screenshots on AHS' press release page.

edit: AHS uploaded a different demo.

 

[Stardust]

This is much better then what I was expecting. Not only is it a update to the engine instead of a whole new engine, but it seems that they will release AI and sample-based VBs at the same for that character time instead of locking them between AI and sample-based based on their upcoming releases This is very good to see.
We also have confirmation that Tsuina is indeed going to be a SynthV VB. I am surprised to see Rikka on CeVIO AI since we do not know the difference yet between CeVIO AI and SynthV AI.

 

[Rylitah]

It hasn't been mentioned in a post yet, so for those who didn't read either press release, Synth V AI is a free update if you own Synth V Studio Pro. If you own Saki, her AI update will also be free from AHS for a limited time.

 

[WyndReed]

I’m really excited about this! (Especially the fact that more software purchases aren’t needed to use it) AI Saki sounds quite a bit better than some of the other AI voices I’ve listened to. Hopefully the promotion with AHS will include users who purchased Saki through Anicute.

 

[Twillby]

Saki AI Lite, a slimmed down voice database will be made available free of charge.

This honestly made me so happy I can't even sjshd (I swear I'm going to buy a bank one day... probably... I just like being able to test them first X'P)

 

[kozet]

I wonder how many CPU cycles the AI synthesis will require compared to the existing ("conventional"?) synthesis. Right now, my laptop is pretty slow at synthesizing voices.

 

[lIlI]

Saki sounds amazing! I'm excited for more AI voicebanks.

Reading between the lines here, it sounds like Volor is at least trying out Eleanor in AI:

 

[uncreepy]

I am worried by the announcement, they only mentioned AHS and didn't mention AniCUTE at all. I bought all my SV2 stuff from AniCUTE!

AHS will offer the AI version for free to current users of Synthesizer V Saki during a limited period.

 

[mahalisyarifuddin]

Coverage from DTM Station:

まるで人のように歌うAI歌声合成の世界がさらに進化。Synthesizer VがAI対応し、従来型とハイブリッドで利用可能に。Sakiユーザーには期間限定無料配布｜DTMステーション

歌声合成の世界において2020年はまさに激変の年。Synthesizer VのWebブラウザ版からスタートし、NEUTRINOの誕生、Synthesizer Vの大幅バージョンアップ、CeVIO AIの発表……と続いてき...

www.dtmstation.com

In the article you can also hear more of Saki AI samples.

And they'll hold a livestream in November 10 discussing about this.

 

[uncreepy]

I was concerned about the lack of mentioning of AniCUTE on the free Saki update and got a reply from a user:

So I went to My Page on AHS and was able to register Saki and the Pro editor.
Here is the quick tutorial I made:

 

[xuu]

Can confirm VOLOR are exploring a variety of options for Eleanor, which is why she might take a little longer. They're really trying their best to make her into the best vocal possible.

Genuinely more excited than I thought I would be to be honest. The AI version sounds so much better than I expected, Saki sounds genuinely human. I'm really looking forward to Rikka's two vocals now and the CeVIO AI Talk vocal.

 

[mahalisyarifuddin]

Tried to remake one of the samples.
https://clyp.it/bxbzmgqx
Because my laptop can't run SynthV Studio I'll just stick with the Web If you have it you can download the SVP in description and compare the samples.

And yeah, she sounds so different yet more beautiful. when I first hear I was like "There's no freakin way this is Saki it's so different! But wait, I love her"

 

[Prism]

It sounds like they captured more of the voice provider. I've always thought sampled based vocal synths like vocaloid always sounded like they are missing a chunk out of the human voice. I'm glad we're are finally able to have a rounder fuller voice