Shadess

Lover of anime, silver-haired waifus and onee-sans. Also a professional cunnysumer. Profile picture by: @UmiHyaku

[FZNUTRI @fznutri]

Fabrics follow a flow and have weight.

[lotovik is making ornament tower 🏰 @lotovik]

close enough! here's how to actually make isometric characters

[pixel_salvaje @Pixel_Salvaje]

That 2-pixel selection perspective trick is a lifesaver!
But if you need to build an entire city or asset pack with it, doing it manually takes forever.
I built PixZels to automate this -it usesmathematical interpolation to instantly project orthogonal views into clean isometric pixel art.

[Melon 🍑 @Melowh_Art2]

Megumin Back and Front 🍑

#konosuba #megumin #brasil

[leanin @prollyleanin0]

One class with rakai later

[Celina 52 Truck Stop @celinatruckstop]

We would like to apologize to this customer this morning after our Gay Manager Jacob briefly forgot he was a homosexual and relapsed into straightness while staring at her for a prolonged period of time.

Jacob converted to being fully gay in 2021, and assured us that this was a one-time incident triggered by his new medication. Jacob is being punished internally.

[りあら.⋆𝜗𝜚 @lial_artistic]

Hair in progress… ⟡.·𝜗𝜚꒷꒦♡°❀⋆.ೃ࿔*:･๋࣭ ⭑

[iisFernandoVol.2 @iisfernando_art]

Helsie our Helsie 🥰

[KuroeMio_枯渺 @KuroeMio_]

想象中的サテラ

[Emerald Apple @AI_EmeraldApple]

This is the future of VTubing.
No more expensive artists and Live 2D models.
No more janky VRoid 3D models.
No more static PNG tubing.

It's a little rough now, but remember this is the worst that it will be, and it will get better, eventually surpassing anything anyone can make with traditional live 2D.

Artists can use this kind of tool to make even better VTubing avatars with smoother transitions between states and better tracking optimizations.

AI models will get better, real-time generative models will get faster, compute gets more robust... and when that happens, all the pixels won't just be recycled states, but new "everything," just like a real person on a camera, except it's all a 2D anime character model.

Or it could be like a model that generates new states, learns from them, and saves them... so it doesn't have to recompute a new state. An ML algorithm will use predictive heuristics just like now to pull the appropriate state automatically that includes prerendered transition animations, kinda like how game character animations are rendered.

Probably be there in only 2-4 years, I bet... maybe even faster.

[Leaflit 🍃 Angel's Sword @Leaflit]

I'm making a program that is a bridge between PNG tubing and Live2d.

It's called "Adventurer". You can make a model in less than 30 minutes and add new emotes in less than 5.

[Ranivan @secondhand9280]

#Emilia #rezero #リゼロ

[Ren @RenegadeRetry]

@Keeen_EN took his shot to impress his oshi @Leaflit and succeeded!

Congrats man!

Other appearances by @TheeHolyBall and @powdury

[rtil @rtil]

Good End

[Lit Silium @LitSilium]

Laundry Day Yor 🩲

Alts on Patr3on

[VIZAZ @_vizaz]

Panty and Stocking 🐰🩷

[にろ┊͙SDミニキャラ @NyoroNiro]

#エイティシックス　バレンタインイメージ　レーナ

[알타리 @Rtari5]

4령사 오버프레임

[Shadess @YamiShadess]

@suitetvapp Does SuiteTV have home screen widgets on Android like TV Time does?? Please say yes🥲