#diffsinger

[aseriesofunfortunatejan]

The [tumblr] vocal synth cover collab premieres on Saturday, April 12th at 10PM (CEST)

Be there or... be square.

29 different people, united by their love for vocal synth and their use of [tumblr], have participated in this cover with a different voicebank each. Led by yours truly, everyone was required to tune their section themselves: making this a truly unique project! ♪(^∇^*)

The description will be updated with full credits after the Premiere - I'm not one to spoil, after all. Open the link above to know what time it corresponds to in your timezone!

[potetopancakes]

100 (+1) vocal synth characters

(suggested by people on twitter)

am considering doing this again with non-vsynth characters, but itll be a lil while before i take on something like that

[utau-ship-polls]

SIDE B ROUND 1

[peeslubn]

Akiyama Daichi / A.I.CHI’s newest Diffsinger model has been released, along with a final demo to showcase his newest language, Polish!!! I hope everyone likes him!

Check out his release video here!

Thumbnail / video art is drawn by @khazure , thank you for working with us!!!!!!

[kiwwiraptor]

~ NEW ORIGINAL SONG ~

『Queen (Girls Like Me)』

Kiwwiraptor feat. 知声 (Chis-A) & LIEE

Production: Kiwwi Lyrics: julieraptor Released: February 19 2025

STREAMING ON ALL PLATFORMS

Lyrics/instrumental/stems available here!

[khazure]

I got to draw Akiyama Daichi for his Diffsinger voicebank release demo!!

[hetatanart]

The full set of Aphelia -stellar evolution- full bodies is done! for now

[runaxee]

NOBODY TOLD ME OF THIS DIVA?!

HAREFORD UR SUCH A DIVA

[gardanana-art]

AI Bomb my beloved!!!

[aseriesofunfortunatejan]

I've said it and I'll say it again - thank you to everyone who participated in this project! The [tumblr] vocal synth cover collab turned out wonderfully!

The premiere went by so fast. Thank you to everyone who attended it! Seeing your hype at the voicebanks, and your enjoyment of everything was so sweet!

(For those who wondered - the countdown theme was the "Dramatic" Theme (aptly named). I selected it because something about it made me laugh out loud when I was cycling through the different themes, but it is good at its job, right?)

As a reminder - everyone tuned their own section themselves! This is unique and very well done. I'm so glad my mixing turned out okay ashdhsha,,

For those who are interested in audio only, or in having a different archive of the project, stay tuned for the SoundCloud release! I'll give it some days, maybe a week, so the YouTube version, video and all, stays exclusive at first. Thank you for your patience!

[pearlstarlight5]

Vocal synth rant from my Twitter

I've realized that the shift from everything being concatenative to everything being AI reminds me of the shift from 2D animation to 3D animation. Yes, a lot of work goes into developing both and AI vocals are much easier to use, but I'm sad that it's like everyone, indie and industry, is abandoning concatenative vocal synths.

I know that Piapro NT will continue to be supported, but once Miku V6 comes out it will be Piapro Who.

And I will likely get comments saying that AI vocals are objectively better, but are they? That's like saying 3D animation is objectively better than 2D animation. It's a matter of taste and what sounds lend themselves better to the music. Even aside from realism, there's still a lot of things that concatenative vocal synths have over AI vocal synths, like reliability, and in my experience, better tolerance to unrealistic emotional tuning and at the same time you can get away with minimal tuning (except for compensating for phonetic glitches in Engloids at least). AI vocal synths tend to sound bored when untuned and respond best to realistic tuning.

Additionally, AI vocal synths are limited by what the human voice is really capable of. DiffSinger best displays this weakness because the voice cuts off after the optimal range. Even auto-tuning in SynthV and CeVIO gets awkward when the range is too high, like for a human. Meanwhile, due to the nature of concatenative synthesis, the vocal synths are capable of a more inhuman vocal range, so they can be used to sing things that humans are incapable of singing. Once again using my experience as an example, my Utau is way better at high notes and low notes than me, and that's the result of her voice being synthesized from a comfortable range rather than my entire range.

[ryouichiii]

Kiiro's 10th Anniversary redesigns!!!

ahhh only 2 months away thats scary ,, i finally got around to finalising his redesigns and im really happy with them!

please look forward to his anniversary on the 30th of November! I have a bunch of covers lined up :]c

[utau-ship-polls]

SIDE B ROUND 2

[peeslubn]

non-human

[shiweimigi]

March 14, 2025 - "I Got No Time" cover song

Cover, mixing: Guillotama

UST: Nukude Choseung

Artwork: @kandavers

Video: @a-purpled-world

Planning: JAE VOCAL PROJECT Team

Vocals: Miyo as Xiao Meihua

Miyo design inspired by @mmayalla 's drawing of Miyo based on Sylvanian Families

Original by The Living Tombstone

Also available on bilibili

[vocaloid-tunes]

Nostalogic (cover) | Keiro Keer feat. Monone // Original | yuukiss feat. MEIKO