
graphics X computers est. 1986 ▲ Art @Meta @RealityLabs ◾️ ex creative partner @SEHSUCHT ⚫️ ex creative director @PSYOP 🌬born at 331.27ppm
Berlin
Joined December 2020
- Tweets1.1K
- Following1.1K
- Followers566
- Likes5.6K
Liebe @bpol_b wieso kreuzt euer heli hier mitten in der Nacht stundenlang rum und weckt alle auf? Danke globe.adsbexchange.com/?icao…
1
81
Mate Steinforth retweeted
Gemini can generate pretty consistent gif animations too:
'Create an animation by generating multiple frames, showing a seed growing into a plant and then blooming into a flower, in a pixel art style'
Mate Steinforth retweeted
Wer war für Sie der beste Kanzler der letzten 50 Jahre? Wer hat Deutschland geprägt und Vertrauen geschaffen? Teilen Sie Ihre Meinung!
Mate Steinforth retweeted
love when @mrdoob spent a decade making shadows, lighting & other shaders simple one-liners and then claude gets all the credit lol
Amazing work here by @robcupisz <3 I adapted a subblock to easily align quads to velocity gist.github.com/MateSteinfor… - should be a built in block in VFX graph! (or is it and Im not aware?)
For the love of our eyeballs, do not align your ember particles to velocity!
Don’t give them that comet texture either, ugh.
Let’s make
✨physically based flying embers✨
A Unity vfx graph implementation and explanation below. 🧵
1
94
Mate Steinforth retweeted
For the love of our eyeballs, do not align your ember particles to velocity!
Don’t give them that comet texture either, ugh.
Let’s make
✨physically based flying embers✨
A Unity vfx graph implementation and explanation below. 🧵
Mate Steinforth retweeted
🐅 Want to rig your favorite meme character?
Try “RigAnything: Template-Free Autoregressive Rigging for Diverse 3D Assets”!
✨RigAnything is a transformer-based model that sequentially generates skeletons without predefined templates. It creates high-quality skeletons for shapes in any pose, completing rigging in under 2 seconds per shape. 🧵(1/n)
Mate Steinforth retweeted
Replying to @leanhealth @GoogleDE
Mate Steinforth retweeted
Maya (the 3D modelling program) was released in 1998 and its advances were generally underappreciated for decades: a rigorously defined scene graph, tracking of edit history (for revising past decisions without undoing subsequent work), data inheritance with change propagation, and a carefully thought-through transformation gizmo. It's only now that teams are trying to standardize around the Maya-influenced Pixar USD format that the industry is coming to appreciate those foundations.
It's interesting to compare to Unreal's innovations back in 1998, which went further in some areas such as a scripting language (UnrealScript) with rigorous forward-compatibility rules, two-dimensional inheritance and change propagation, automatic networking, and a similar metadata-centric design in which script declarations drove both code and an artist-centric property editing user interface.
Maya and Unreal came soon after Java (1996) and soon before C# (2000), and provided an engine-specific analog of many of their language innovations. Alas all four systems, caught up in product release cycles and immediate customer demands, settled for incremental improvements in data architecture instead of the dramatic advances the architects foresaw in the early days.
I feel the best version of a 3D architecture is still to come, and isn't far away, with the frontier moving to collaborative editing in a revision-controlled environment; rigorous server-to-server and client-to-server networking enabling unlimited scalability; more fully-featured inheritance, change propagation, and edit history; programming language advancements towards expressive power and rigor; and a standards-worthy definition of the whole stack, analogous to web standards.
en.wikipedia.org/wiki/Autode…
anyone knows of a distortion shader (read render feature) that includes transparent objects, works in unity 2022.3 LTE, URP, VR, single pass and conforms to conventions (RTHandles etc) and doesn’t drive the user insane? :) cc @ushadersbible
1
106
Mate Steinforth retweeted
Took us a while but looks like the @threejs community finally caught up to @googleearth 👏
I like the artistic control three-geospatial clouds are allowing for in a 3d-tiles-renderer world. It's working quite well for creating realistic fog effects. Thanks to @shotamatsuda and @garrettkjohnson for two really great packages! Now I just need to add water..
Mate Steinforth retweeted
🌐 research.nvidia.com/labs/tor…
The inverse renderer is a video diffusion model conditioned on input video and domain embeddings, producing accurate, temporally consistent scene attributes such as normals, albedo, roughness, and metallicity.
Hey @BVG_Kampagne @BVG_Ubahn why not prefill the point of departure with the nearest/last used station and especially why not prefilled number of tickets with 1? Who wants to buy 0 tickets? Optimize for least clicks 💖
3
185


















