Saki AI my beloved, she will always sound good, no matter what you do, a speaking test like this, or a song thats hard to tune, or even if you are new into tuning and don't know how to use her, she will always find a way to randomly sound natural and realistic, ahdkbskfbsn, okay, enough Saki simping, i love how she sounds in here, and love the other dialogues as well, but Saki was def my favorite
While this is amazing, I still feel some improvement can be done. Is there a way to either get more data points or remove some so it's somewhat smoother?
Can you do O Come all ye faithful Japanese version please? Here's the lyrics. ruclips.net/video/CoKwDLpY1uM/видео.html And Here's the vocals. ruclips.net/video/fOmHXeYhDU8/видео.html Also, you can release a SVP file in your Google drive so everyone can give you credit.
It depends on if i am able to separate the vocal in a good quality, it is harder when there are multiple voices or effects. i can release the SVPs - any particular you want? but there is a lot of note bending - it is not much useful as a midi for playing
![Comparing Female SynthV AI Voicebanks [Washing Machine Heart - Mitski]](http://i.ytimg.com/vi/aKvoRLR7l78/mqdefault.jpg)
SAKI sounds the closest to original and her tone makes it sounds authentic and less synthetic
agree & wonder how it sounds to japanese listeners
Maki sounds very natural though too, but more anime voice acting like..
I think it's bcs Shimakura Rika (the original) is voice acting
Saki AI my beloved, she will always sound good, no matter what you do, a speaking test like this, or a song thats hard to tune, or even if you are new into tuning and don't know how to use her, she will always find a way to randomly sound natural and realistic, ahdkbskfbsn, okay, enough Saki simping, i love how she sounds in here, and love the other dialogues as well, but Saki was def my favorite
Saki sounds real and closest to tone but she sounds a bit shaky
interesting!! they actually sound really good
Tsuina and Maki sound like anime lines x)
programming synth v for speech must be really hard
With the Real Voice scripts it's the same as for singing.
While this is amazing, I still feel some improvement can be done. Is there a way to either get more data points or remove some so it's somewhat smoother?
It was tuned with Rikka, then I only switched other voice banks. But I agree, it could be better.
HOW DID YOU GET GENBU TO SOUND SO GOOD
Copying a real voice makes the difference & the lowest note should be at least C3 - lower is not good
Wuau
Wow
HOW????
ruclips.net/video/XTuHtN9h6Ak/видео.html
and
ruclips.net/video/QmdrbOsF_7Q/видео.html
Can you do O Come all ye faithful Japanese version please? Here's the lyrics. ruclips.net/video/CoKwDLpY1uM/видео.html
And Here's the vocals. ruclips.net/video/fOmHXeYhDU8/видео.html
Also, you can release a SVP file in your Google drive so everyone can give you credit.
It depends on if i am able to separate the vocal in a good quality, it is harder when there are multiple voices or effects. i can release the SVPs - any particular you want? but there is a lot of note bending - it is not much useful as a midi for playing