Menu
What's new
By:
Filters

SynthV Solaria is UNREAL

SamTheAudiocrat said:
Here it is: https://dreamtonics.com/synthesizerv/

I am using the Solaria voice model. I want to get the Kevin voice model as well to compliment the female vocal. It does come with a free voice model called Mai. Mai can sing both Japanese and English, but has a slight accent when singing English, which is why I bought Solaria as well. Will pick up Kevin later as I think Kevin has a nice contemporary male voice that can do both pop/rock and cinematic. Would be lovely to layer two voices and get a duet.
Click to expand...
Word of warning - Kevin is very "square" compared to Solaria. I'm not saying its hard to get him to do a realistic performance, but his style is very straight and requires more work than I think any of the other voices, in order to make him sound relaxed.
 
SamTheAudiocrat said:
Just bought SynthV and got the Solaria voice because that it'll perfectly fit Epic Cinematic Vocals. Sat around today for 30 minutes to test it out and darn, I'm super impressed. I've heard a lot of demos online and sometimes you can make out it's AI generated, but it believe if you put in some time and some thought into how a real human sings and emulate that through your note and syllable selection you can make it sound very very convincing. Here's a quick example of what I could get:

View attachment 10,000 Candles in the Wind.mp3

Would love to hear your thoughts and even a few demos if you've tried this out with Synth V.
Click to expand...
Every now and then you click on a mockup or composition, and it leaves a mark.

You had to go and remind me of 'lil Sebastian. Now I'm crying in the corner of the room.
 
Aymara said:
But you can use several databases in one project to create a choir ;) Alternatively you can use the same voice on several tracks in a project.
Click to expand...
Yep! That's exactly what I've been working on. A client is having me produce the music for his upcoming album, but hiring a real choir is not in the budget. SynthV to the rescue. I built these 8-piece choirs out of Solaria, Weina, Mai, Kevin, and Frimomen.

What you're hearing are all works in progress — unmixed, no lead vocals, lots of blocked-in placeholder instruments, and plenty more choir editing to take care of. In the final version, the choirs will be topped off with two real singers.

View attachment choirsamples.mp3
 
fieldmousetrebuchet said:
The big problem with using it in final works is lack of expressiveness ...
Click to expand...
It requires fine tuning ... check the Youtube Channel I posted above. Ok, there are things, that are difficult like whispering voices or impossible like screams, but it can be more than just being pretty realistic.
 
SamTheAudiocrat said:
Haha. I tried again with Solaria and pitching the vocals higher only makes her scream. Is there a particular range for the falsettos?
Click to expand...
I don't know actually - I really did the absolute minimum on this! I just ran an acapella version of the original track through Synth V's Audio-to-MIDI feature, corrected a few words and voila! The project idea was never pursued due to copyright concerns, so I never got round to fully rewriting the song.
 
SamTheAudiocrat said:
Haha. I tried again with Solaria and pitching the vocals higher only makes her scream. Is there a particular range for the falsettos?
Click to expand...
For Solaria, it sounds to me like it starts around B4. Try it with the soft vocal mode, or increase the breathiness in her voice. I think it does depend to an extent what she's singing.
 
richiebee said:
Word of warning - Kevin is very "square" compared to Solaria. I'm not saying its hard to get him to do a realistic performance, but his style is very straight and requires more work than I think any of the other voices, in order to make him sound relaxed.
Click to expand...
Hey I finally went ahead and got Saros instead because I already loved Solaria from the same folks. Saros is quite good and serves the purpose too. Looking for more such voices from Eclipsed Sounds 😃
 
Aymara said:
Just some inspiration ... just to show, what's possible:

Click to expand...

I'm kind of glad that it sounds really lifeless compared with the original (the vocals had a lower volume than what's normal for this kind of music) – but that would possibly be true if a number of human singers tried as well.

If use of AI based vocals will be common, for instance within 10 years, that would feel like a major step in the wrong direction for me. I like to know that what I'm listening to is a real singer, and not a robot.
 
Vik said:
I'm kind of glad that it sounds really lifeless compared with the original (the vocals had a lower volume than what's normal for this kind of music) – but that would possibly be true if a number of human singers tried as well.

If use of AI based vocals will be common, for instance within 10 years, that would feel like a major step in the wrong direction for me. I like to know that what I'm listening to is a real singer, and not a robot.
Click to expand...
i used SynthV for my latest metal album.. it was a huge step forward for me since i never had "access" to any real singer and for an hobby it would have been too difficult to find someone to help me to sing my song.

regarding your comment... 10 years ago orchestral sampling was maybe very irrealistic and you probably would have spotted fake samples. now orchestral samples are very realistic, at least for the "common listener". if the evolution of synth voices will be the same, or faster, we can expect a huge step forward in expressiveness and other nuances... who knows what will be in 10 years!
 
winn3r said:
I wasn’t a fan of these until I heard this one.. unreal.


The English pronunciations are sometimes slightly off, but good enough for sketching ideas before they go to a human singer (or not). I also really like Gumi/Megpoid - sounds a lot like one of my Japanese friends :laugh:
Click to expand...

That's the best megpoid demo I've heard yet. Maybe I'll pick her up because DL site has her on sale. Really, most of the megpoid demos she sounds really robotic to me. That was pretty good.

Synth V's mandarin has been really good in general across most of the voices. I'm most impressed with that. Then the english and then the japanese - but only because I'm not as as fluent in Japanese as I am in Mandarin so I'm not 100% on how it sounds. Mandarin I got a degree in, back in the 90s. Japanese I Learned because in the mid- late 90s I lived in Tokyo and worked on anime songs and jpop for a lot of companies. I worked on music for macross, etc. So unfortunately I never got a school education in Japanese. I should just go take some classes on it. Most of my Japanese is just picked up from sitting all day everyday in studios with jpop singers lol.

No comment on the cantonese though because I don't speak it so I'm not sure.

Minus sounds great but I don't hold my breath for her because buying off taobao is expensive and a massive pain in the butt for anyone outside of china. Shame because she does sound good, along with Stardust, but I'm not holding my breath on getting either anytime soon. Shame they're limiting such good voicebanks to such a small audience.
 
OPY said:
If she could sing my native language (Finnish), I would gladly let her do it, also. But that I must do, at least for the time being, myself... :emoji_scream:
Click to expand...
You can make synth v sing in finnish but it takes a bit of manual work and there is a slight accent
 
  • Like
Reactions: OPY
Vik said:
I'm kind of glad that it sounds really lifeless compared with the original (the vocals had a lower volume than what's normal for this kind of music) – but that would possibly be true if a number of human singers tried as well.

If use of AI based vocals will be common, for instance within 10 years, that would feel like a major step in the wrong direction for me. I like to know that what I'm listening to is a real singer, and not a robot.
Click to expand...
Synth-V has given me a purpose to write songs again.

When I submitted this one as part of my RPM in Feb, I had people debating whether it was a real singer and I had multiple people thinking that it was a real singer, commenting on the beautiful voice.

View attachment RPM-TheSunComesCloser-POST.mp3

I used to work with real singers, but being completely unable to sing my self, or really well articulate what I wanted them to change from my lead sheet, it was always an exercise in frustration, and took way longer than it should, with multiple vocal sessions in order to get a song into the shape I wanted. While I realize that's my own failing, Synth-V has changed it for me.
 
You must log in or register to reply here.
Top Bottom