SynthV Synthesizer V Discussion Thread

[kylin]

I stand corrected! Looked into the whole singing mode shenanigans and turns out you can set manual as your default preference.
Here's a basic guide and overview for those wish to have manual as their default:

Sing Pitch Mode - Synthesizer V Studio Unofficial User Manual

The unofficial user manual for SynthV Studio

manual.synthv.info

So I guess for the new update, I have no complaints!

 

[morrysillusion]

thought someone wouldve posted this.

D-Lin and Ritchy have both had their FULL releases. Their feature limited trials will still exist though, it seems.



i commented this on Ritchy's video but its a shame it looks like a Dreamtonics person just tuned him, yet D-Lin gets tuned by Creuzer?? they both sound good but i know their english vocals would sound even better if they got a tuner like that to do their demos...

 

[morrysillusion]

making a seperate post to point this out since it was kind of important, hope thats ok. i dont think even mentioned in those tweets/videos, so

They both have vocal modes now.

for both:
HighTone
MidTone
LowTone
Singing

Was just thinking about how much i still liked Ritchy's singing voice, and now it looks like you can properly put them in a more "singing" mode properly, which is sweet

 

[morrysillusion]

thought someone would have shared this here!! was gone all day but now i like it


i know many werent super interested in HDVM when it came out-- i definitely had my uses for it, but forgot it existed in many cases. even though i think it would be useful in many ways (i have many times where i know i need something different or an alt take) i just kind of forget its there.

but this seems interesting-- seeing the changes actually affecting your default pitch generation... as i understand, RLHF will basically learn off of your retake choices...? it seems that it learns based off what you stick with i guess. not sure if i am getting that right. also not sure if there will be an 'off' button in a sense (they have provided ways to change things from being the Default so maybe) but also if you just... dont use the retakes feature it wont matter.

i do think this is cool though. it really hit me on Ryo tbh. i have Ryo but i havent really been a fan of him... his self tuning especially. so seeing how it changed in this example was actually something i really wanted to see for the voices i tend to use their self tuning for (sometimes i generate that, then go back to manual so it makes the pitch lines permanent, and work off it from there!) honestly in the end i think stuff like this always applies less to the older school vocsynth users who are just used to tuning from scratch-- i think this will definitely benefit the users who dont, especially the ones in the professional/non vocsynth spaces who well. dont tune like, at all.

 

[lIlI]

Synthesizer V 1.10.0 features have been announced, seems like some simple quality of life adjustments.

Copy and pasted from the description:
Feature 1: This new update utilizes Reinforcement Learning from Human Feedback (RLHF), a method previously employed in text generation AIs to steer the system towards generating more preferred responses. With this technology, which we are the first to apply to audio, we can "instruct" the Synthesizer V AI to generate pitch curves that users find appealing. This translates to fewer off-key errors, better use of vibratos, and more expressive transitions between notes.

Feature 2: We are enabling feedback submission for all users, to improve pitch generation and extend RLHF to the modeling of voice timbre. Within the AI Retakes panel, the heart icon on each take now also serves as a button for users to share their preferred takes with Dreamtonics. Models trained with this feedback data will be deployed in future voice database updates.

Feature 3: This update incorporates several usability enhancements. A track color selector is now integrated into the right-click menu on track headers, providing an efficient method for track differentiation and organization.

From their detailed update notes:

• Plugin: AU and VST3 plugins now support drag-and-drop importing of MIDI files.
• File: disabled forward-compatibility for projects saved in Synthesizer V Studio 1.10.0 and onwards due to a breaking change in the value range of take Expressiveness. Users may use “File/Save As (1.9.0-Compatible)…” to export the project to an earlier version of Synthesizer V Studio.
• Synthesizer V Engine: fixed crashes and quality degradation on standard voice databases due to some dictionary changes introduced in the last update.
• GUI: fixed missing fonts on some keyboard shortcuts on macOS.
• Live Render: fixed a bug causing the renderer to skip some parts of a project when loaded immediately after startup.

 

[PearlStarLight5]

For her 5th anniversary, Eleanor Forte now has an official Twitter https://twitter.com/Eleanor_Off

 

[MillyAqualine]

Ah ? I thought she used to already have one.... Seems I've mistook with another character then

 

[poser]

 

[lIlI]

A second beta for 1.10 has been released (for manual download only)


Download

 

[mary34]

A quick test i did on one of my project files with the new beta

They seem to have fixed the weird vocal mode bug where some of the vocal modes did not make much of a difference
Cross lingual also received further enhancement to pronunciation, Teto is the most noticable improvement but cross lingual sounds moren natural across the board now (I don't have a preview of that rn on hand)

 

[morrysillusion]

wanted to jump in with some comparisons too. a little conflicted, this example is a song i have jumped into over and over since 1.6 and 1.9 actually gave it the power i wanted that was more comparable to the original singer? so its interesting to see how different it is. but, 1.9 was literally peaking on default too lol. but these files are not changed at all, the very end has power and solid at 100%

Solaria with english. there are also some other little bugs i hear too that i will try to report.
here is 1.9

Solar 1.9 ver_masin

Listen to a track by morrysillusion.

www.whyp.it

here is 1.10.0b1

Solar 1.10 beta1 ver_masin

Listen to a track by morrysillusion.

www.whyp.it

here is 1.10.0b2

Solar 1.10 beta2 ver_masin

Listen to a track by morrysillusion.

www.whyp.it

and of course a comparison of the language improvements, here with Yuma 1.9 and 1.10.0b2

yuma 1.9 1.10 compare

Listen to a track by morrysillusion.

www.whyp.it

this one also shows that tone change where they sounded stronger-- yuma was on 75% solid and 35% power. which i can definitely say Yuma was waay too strong, it resulted in some weird quirks like his voices getting a but gravelly on higher, long notes. (admittedly it was cool at times, which having multiple versions makes this nice). so the fix is understandable, but in many ways it kinda mellowed them out a lot more than i expected. i recall then trying to make these vocal modes *more apparent* which felt obvious in 1.9, but now it kinda went backwards from that? but they do sound more differentiated which was also their attempt in this version.

EDIT forgot to say for the Yuma one- it uses Synth v's generated tuning and for 1.10 it uses the 'feedback' so it is different than 1.9. i also am still learning that new feature so the tuning may sound a bit off in places as i try to give it more feedback

 

[Mika]

The demos for 4 (!!) new Dreamtonics voices are out! Wei Shu (Cantonese Male voice), Sheena (Japanese/English Female voice), Eri (Japanese Female voice) and Hayden (English Male voice).

 

[Leon]

Whoa, 4! They all sound so nice too. I think my favorite of these four might be Wei Shu. Eri is also pretty neat, but RIP Haru I guess won't be the first JP rap aimed vocal anymore pfft

I know people don't care for the alphabet soup's tacky letters, and same, but the banks themselves are always so high quality. Chefs kiss

 

[Blue Of Mind]

I think Wei Shu might be my current favourite as well, but we all know Dreamtonics always undersell their VBs when it comes to demos. I'd like to hear more examples of Sheena in particular, because hell yeah, bilingual VB without using the crosslang function.

 

[mary34]

Whoa, 4! They all sound so nice too. I think my favorite of these four might be Wei Shu. Eri is also pretty neat, but RIP Haru I guess won't be the first JP rap aimed vocal anymore pfft

I know people don't care for the alphabet soup's tacky letters, and same, but the banks themselves are always so high quality. Chefs kiss

That depends on when Eri actually releases. I think Haru is supposed to be release some time this november if I recall correctly? DT tends to take a month or two between announcements and release with their first party banks so from what we know Eri won't drop until after Haru. It would be lowkey funny tho if she drops first :D

 

[Cerāgi <º•]

That depends on when Eri actually releases. I think Haru is supposed to be release some time this november if I recall correctly? DT tends to take a month or two between announcements and release with their first party banks so from what we know Eri won't drop until after Haru. It would be lowkey funny tho if she drops first :D

Poor Haru ;;... oh well, at least she will have a rap companion by then :3

 

[lIlI]

I hope they drop together, and we get a vocal synth rap battle!

 

[PearlStarLight5]

Someone make them do the Zombieland Saga rap battle!

 

[morrysillusion]

oh what a surprise! i like the general sound of Sheen a lot, though some enunciation isnt my style. i actually like Hayden a lot tbh, but i wasnt super into Kevin's tone and prefer this kinda sound for a male. i think im prob gonna get him since there arent many native english males that i have liked. wei shu also sounds great! very strong.

edit: also dont know if its just the mix bringing it out but Hayden handles some S's and th's like yuma a bit? kinda like when yuma had a sorta lisp? not that i care tbh, will have to see if thats a consistent thing or just mix related

 

[AmazingStrange39]

oh what a surprise! i like the general sound of Sheen a lot, though some enunciation isnt my style. i actually like Hayden a lot tbh, but i wasnt super into Kevin's tone and prefer this kinda sound for a male. i think im prob gonna get him since there arent many native english males that i have liked. wei shu also sounds great! very strong.

edit: also dont know if its just the mix bringing it out but Hayden handles some S's and th's like yuma a bit? kinda like when yuma had a sorta lisp? not that i care tbh, will have to see if thats a consistent thing or just mix related

I do notice a bit of a lisp...honestly that means he reminds me even MORE of Avanna than he did before

this is like...avanna fans' dream male vocalsynth (referring more to her pop leaning and general tone than her celtic-ness)