TLDW: Cantonese and rap support are in the works
SynthV Synthesizer V Discussion Thread
- Thread starter xuu
- Start date
-
- Tags
- synthesizer v
TLDW: Cantonese and rap support are in the works
Vocaloid not too long ago: *adds Mandarin cross-lingual*
SynthV: *adding a 4th language cross-lingual*
CeVIO: *releasing Haru in a couple months and will somehow incorporate rap*
SynthV: *adding a rap update*
Maybe Cantonese has been in development a while, but I feel the need to point this out in case anyone else wants to get into conspiracy mode.
SynthV: *adding a 4th language cross-lingual*
CeVIO: *releasing Haru in a couple months and will somehow incorporate rap*
SynthV: *adding a rap update*
Maybe Cantonese has been in development a while, but I feel the need to point this out in case anyone else wants to get into conspiracy mode.
very cool to hear the rap incorporation! i feel likes rap style singing has been greatly ignored in vocal synth (maybe part in due to the struggle to make it sound good).
also its no surprise at all theres a pattern of 'company adds new thing, other company adds new thing (and sometimes more)'. vocaloid 6 coming out with the similar features as synth v *and also cross language seamlessly*. synth v then making their seamless as well. none of these things are patented so its free game for everyone to keep adding the cool new things when someone else does it! because competition.
also its no surprise at all theres a pattern of 'company adds new thing, other company adds new thing (and sometimes more)'. vocaloid 6 coming out with the similar features as synth v *and also cross language seamlessly*. synth v then making their seamless as well. none of these things are patented so its free game for everyone to keep adding the cool new things when someone else does it! because competition.
It's definitely good that engines are basically bouncing off of each other! Results in better product for all of us 
The only toxic aspect is when people get into heated debate about who is the ""top dog"" ( not at all accusing Pearl of doing so just to clarify, you're a cool bee Pearl 
)
The only toxic aspect is when people get into heated debate about who is the ""top dog"" ( not at all accusing Pearl of doing so just to clarify, you're a cool bee Pearl )
the synthesizer v machine generating new features:
More language support is always great. The voice in the native Cantonese clip is nice.
Can't make any judgement calls about the quality of the synthesis, don't know any canto. But from an untrained ear, it sounds cool.
The rap integration... I'm really happy that they're trying this.
I'm personally not absolutely blown away by the short clip provided in the trailer, but I am feeling like the target audience for this future is probably C-POP drawing from the K-POP explosion (and this is also making me feel like Korean might be the next language to be supported). The rapping demonstrated is just so reminiscent of some of the stuff I've heard in passing in K-POP that additionally disseminated to C-POP instead of being directly inspired by the source genre. Musically it's, uh, interesting. Not trying to say I'm trying to put a ton of stock in this tech demo, but it does make me wonder how it will function in software, and how flexible it will be when it comes to expression.
I am expecting it to be very model-dependent, and on that basis, I wonder what that model will have been trained on in the end.
I'm excited to hear the Japanese version, though. I wonder if it'll be able to do classic 90's inspired stuff like SOUL'd OUT, or if it'll be more restricted to "rap parts" in Japanese songs generally (meaning the enunciation isn't really stylized, but the focus on the vocalization moved to the rhythm), if it'll be like how utaites rap (lol, I wouldn't be against it) or if it'll be all of the above.
Really intrigued to see how it turns out on a technical basis. If rapping is on the table on an engine level, I wonder what other things are, too?
More language support is always great. The voice in the native Cantonese clip is nice.
Can't make any judgement calls about the quality of the synthesis, don't know any canto. But from an untrained ear, it sounds cool.
The rap integration... I'm really happy that they're trying this.
I'm personally not absolutely blown away by the short clip provided in the trailer, but I am feeling like the target audience for this future is probably C-POP drawing from the K-POP explosion (and this is also making me feel like Korean might be the next language to be supported). The rapping demonstrated is just so reminiscent of some of the stuff I've heard in passing in K-POP that additionally disseminated to C-POP instead of being directly inspired by the source genre. Musically it's, uh, interesting. Not trying to say I'm trying to put a ton of stock in this tech demo, but it does make me wonder how it will function in software, and how flexible it will be when it comes to expression.
I am expecting it to be very model-dependent, and on that basis, I wonder what that model will have been trained on in the end.
I'm excited to hear the Japanese version, though. I wonder if it'll be able to do classic 90's inspired stuff like SOUL'd OUT, or if it'll be more restricted to "rap parts" in Japanese songs generally (meaning the enunciation isn't really stylized, but the focus on the vocalization moved to the rhythm), if it'll be like how utaites rap (lol, I wouldn't be against it) or if it'll be all of the above.
Really intrigued to see how it turns out on a technical basis. If rapping is on the table on an engine level, I wonder what other things are, too?
Last edited:
Sorry I ROFL'ed hard at this video x'D
:rofl: (Also it didn't help because I kept watching walkthroughs lately) Then regarding the demo, it sounds fairly nice but yeah you can tell it's just starting for the rap
Some songs I've been wanting to cover have rap parts, and getting it to sound good can be bit of a trial, so I'm really excited to see rap support being added. :D Now all they need to add is the ability to choose/change track colors (I sort my tracks by color) and glottal effects (growls, vocal fry, etc.), and I'll be content lol.
If you have either vocal by Eclipsed Sounds, they provide a script for changing track color
So Mais voice actress was indirectly "announced" by her posting a cover of Mai's demo song on her YouTube channel. They sound almost exactly alike 
Her name is Honoka Kitta (橘田 ほのか) and she's a singer-songwriter with 6 albums
Her website: DISCOGRAPHY
Her name is Honoka Kitta (橘田 ほのか) and she's a singer-songwriter with 6 albums Her website: DISCOGRAPHY
directly now too!
Oh she sounds gorgeous. Mai is by far my fave out of the SynthV family so it’s cool to hear her VP as well!
Oscar AI was properly confirmed to be a SynthV (voicemith ily but so many of your announcement posts are so ambiguous)
Synthesizer V Studio 1.9.0b1 Update: Rap, Cantonese and More | Dreamtonics株式会社
Dreamtonics releases Synthesizer V Studio 1.9.0b1 update. As a major update to Synthesizer V Studio, version 1.9.0b1 is being released as a beta update. This new version introduces Rap Vocal Synthesis
1.9.0 Beta has been released, including Cantonese support and rap as some main updates. Their two new rap vocals, Ritchy and D-Lin, are also available to download as a Feature-Limited Trial (FLT) ahead of release.
Ritchy's voice is based on B-Bandj. (Source: DT's Tweet)
Gave it a quick 5-10minute go and it's definitely fun, but I have some gripes mostly with how it messes with how I personally use the editor. "Sing" mode seems to be the default now, which is basically Instant Mode which, whilst it has improved, I'd still rather it be an option I can opt into using rather than having to opt out via selecting all notes, choosing "Manual" and then clearing all pitch edits it made before doing so. I'd like an option in settings to choose the default or something, if possible. (Also I'd like if the tracks in the arrangement panel would change note colours to match the notes on the piano roll, instead of staying green.)
The Rap function is really simple to use from just fiddling around, and with it just being a simple parameter where you click and drag arrows it's quick to pick up, but may take longer to get like really snazzy results (Mandarin Chinese has some extra functions too, but I don't feel like I'm best for judging the rap functions in a language I'm not too familiar with). I'm really into Ritchy's accent though, even when used for singing instead of rapping. It definitely stands out.
Overall a lot of the new functions aren't something I'd use extensively, but just having the option to use them is nice.
Last edited:
Listening to some tests with the new rap feature, it sounds like it will also be good for shouts and chants; this should add a lot to what producers can create!
This is really cool! And as I was telling a couple of friends before, I think that a rap vocal/function will be really helpful even if you don't listen to much "rap music" because a ton of songs have bits of talk-singing that took a lot of work to replicate, usually with slightly awkward results. Even for things like musicals now, which often feature bursts of talking or some vaguely-sung shouts, etc. It's a very nice extra option that I look forward to messing around with.
looks like the 2 new vbs don't even have letter logos. Are letters too anime nowadays?
I was thinking the rap function could potentially improve the quality of Talkloid videos! It will be interesting to see experimentation with the rap feature beyond its intended purpose.
I really like the way they decided to implement the feature. UX-wise, looks great and musically, the "arrows" visualization is intuitive.
They must be proud of it, given the pre-release trial for the two voices-- unless they've done this before?
I bet they just haven’t designed logos yet.
As far as this feature's ability to genuinely represent the various styles represented by R&B and rap, I still have doubts, but also high hopes.
Everyone is right that the opportunity for shout/chant/talk parts is absolutely there. It makes me wonder what third party support for this feature will be like. Given that it relies on... what looks like an almost completely separate dataset? I wonder if it'll be widely accepted outside of the Dreamtonics voices. After all, hiring someone that can rap and sing is often an entirely different ballgame than hiring for one or the other.
That prospect does bum me out slightly given that some of the voices I want to hear with this feature the most are the moe-type voices created by 3rd party developers. There's a lot of vocarap out there.
But we'll just have to wait and see. Maybe Dreamtonics will create an incentive for developers to update their vocals-- or maybe they won't and "rap vocals" will become their own niche on the engine.
I am excited to see HARU release and be able to compare the range of a CeVIO vocal with specialized rap data/features and SynthesizerV's features.
They must be proud of it, given the pre-release trial for the two voices-- unless they've done this before?
I bet they just haven’t designed logos yet.
As far as this feature's ability to genuinely represent the various styles represented by R&B and rap, I still have doubts, but also high hopes.
Everyone is right that the opportunity for shout/chant/talk parts is absolutely there. It makes me wonder what third party support for this feature will be like. Given that it relies on... what looks like an almost completely separate dataset? I wonder if it'll be widely accepted outside of the Dreamtonics voices. After all, hiring someone that can rap and sing is often an entirely different ballgame than hiring for one or the other.
That prospect does bum me out slightly given that some of the voices I want to hear with this feature the most are the moe-type voices created by 3rd party developers. There's a lot of vocarap out there.
But we'll just have to wait and see. Maybe Dreamtonics will create an incentive for developers to update their vocals-- or maybe they won't and "rap vocals" will become their own niche on the engine.
I am excited to see HARU release and be able to compare the range of a CeVIO vocal with specialized rap data/features and SynthesizerV's features.
LOL. I won't claim to know for sure, but I 100% assumed it was just that they were free trials and their full releases will have actual logos