Bart Wronski 🇺🇦🇵🇸
@bartwr.bsky.social
Engineering, Computer Graphics, Art, DSP, ML Culture, Techno, Industrial, and Electronic Music. Research Scientist at NVIDIA. Ex Google Research, Ex games (Sony, Ubisoft, CD Projekt). Politically leftist. He/they. https://linktr.ee/bartwronsk
Two cheap "gadgets" that are *dramatic* quality-of-life improvements: - General, everyone: USB type cable tester, both data + power (optionally extra power-only one). Saves sooo much time and frustration (which USB is that? Power only, how much W, or data?). - More "IT" folks: a cheap remote KVM.
My favorite paper from SIGGRAPH this year was "Taming Optimization Variance in Compact Neural Shading Networks" by Bitterli et al; I think this paper will be seen as a really influential paper over the next several years. (1/3) research.nvidia.com/labs/rtr/pub...
Taming optimization variance in compact neural shading networks | NVIDIA Real-Time Graphics Research
<p>We present a training algorithm to mitigate optimization instabilities in small neural networks, like those used in real-time neural shading applications. While large, overparameterized models exhi...
research.nvidia.com
Hopefully one last SIGGRAPH advertisement. I, myself, yours truly, will present on Haolin's behalf a data-driven method for joint adaptive sampling and quadrature for integration. Wed R408A. In the end of the talk I'll argue why everyone is doing the wrong thing. Paper: suikasibyl.github.io/nqr
I published vibrance-0.1.0 on Maven, now with support for Compose for Desktop (JVM). The APIs let you mix colors yourself, or draw gradients as modifiers or Compose brushes. Next steps: refine the model training to improve some color transitions, better alpha support. github.com/romainguy/vi...
We actually get to witness Nvidia open sourcing their Neural Appearance Models in this lifetime. Thanks for finally dropping the guesswork reimplementing all of that. github.com/NVlabs/neura... Their new SIGGRAPH is also out: research.nvidia.com/labs/rtr/pub...
GitHub - NVlabs/neuralappearance
Contribute to NVlabs/neuralappearance development by creating an account on GitHub.
github.com
Ugh I don't comment on gamedev news. Not because I don't care or empathize, but because it's an equivalent of the shitty "thoughts and prayers." Doesn't change anything, powerlessness probably depressing for the affected/victims. But the recent days/weeks/months... Holy shit. Capitalism sucks.
bun.com/blog/bun-in-... incredible visualizations and a great in-depth description of the agentic (terrifying! ;) ) automatic porting workflow. Agentic coding might be a true opportunity for Rust. I don't want to ever solve its puzzles. But maybe I won't have to?
Rewriting Bun in Rust
Why & how we rewrote Bun from Zig to Rust
bun.com
my #1 grizzly beard protip for (game) dev is: if you have a bug that you dont understand the cause of immediately, dont try to fix it; add & iterate visualization / inspection tools until the cause is plain to see. the accumulation of 'observability' in your system's state will pay off many times
god i hate programming. always have. always will. damn if you can't do some neat stuff with it though.
personally i'm ok with AI techniques being less well known but there's a deeper thing going on here which is far more important IMO, because it's also partially why LLMs have taken over == this thread is in response to this tweet: == x.com/krismicinski...
Implicit surfaces are everywhere in vision and graphics applications that optimize shapes. If we adjust the surfaces through a "learnable" deformation field, how can we ensure they stay close to the undeformed state (e.g. a template)? Our latest paper offers a simple and effective ARAP regularizer🧵
The code, including handle deformation and stylization demos, is available here: gitlab.com/tobidju/arap... Or, just run `pip install iarap` if you only want to use the regularization term. Thanks to @tobidju.bsky.social for leading the project + coauthors @ugoalone.bsky.social and Marc Alexa.
Tobias Djuren / ARAP-Regularization-for-Implicit-Surfaces · GitLab
Source code for the paper 'As-Rigid-As-Possible Regularization for Implicit Surfaces' at SGP 2026.
gitlab.com
Scapegoating AI (HBMs!) for the DRAM prices is so silly and the "same old" story over again. Corps and price fixing cartels getting away with their stuff thanks to "useful idiots" (sanctimonious luddites). Real systemic issues -> convenient clickbaity scapegoat -> we'll never get a systemic change.
lex luddy@lexluddy.xyz · last mo.Reminder that memory manufacturers (including Samsung, Hynix, and Micron) were found to be an actual dictionary definition "cartel" back in the 2000s and were accused of illegal price fixing again in the late 2010s. They have been trying to screw you for literal decades.
*tap tap* Is this thing on? www.youtube.com/watch?v=B9bz... rookandpossum.com/posts/scanli... I made a thing! I suspect I'll be posting to an empty room like a madman, but for now, I'm going to make a habit to start posting things here when I post videos/blogs/papers.
The Scanline Sweeper: A Glyph Rendering Algorithm
YouTube video by Rook and Possum
youtube.com
A new article: Revisiting Yliluoma’s ordered dither algorithm 30fps.net/pages/revisi... It's a deep dive to a single "color selection algorithm" used for ordered dithering. Nothing groundbreaking but I did uncover a neat simplified version that works OK. It's technical but with visual examples.
I made a video about the smooth-maximum function, I hope you like it! youtu.be/6Qb6QtC6QMs
A function you cannot do without
YouTube video by Inigo Quilez
youtu.be
Our #SIGGRAPH2026 paper "Multi-Layer Reservoir Splatting for Temporal Reuse under Disocclusion" improves #ReSTIR by reusing previously occluded samples. Video: youtu.be/-ya1bBnwx38 Project Page: graphics.cs.utah.edu/research/pro...
Made a tiny color lib: NuteLch 🌰 It sits between OKLCh and OkHSL. You dial chroma as a fraction of the way to the gamut cusp (so "saturation" means the same thing at every L & H, never out of gamut), but the result is still a plain oklch() color. meodai.github.io/nutelch/
AI: virtually no subject provokes such vitriol in my communities. I can grant the haters many of its harms, but their solutions leave much to be desired. I fear for the future even as I wonder at the future we are living. And so, dear hater, I must write: shaping.systems/blog/ai-suck...
AI sucks. Hating it is not enough.
A jeremiad on uninvention and bitter hope
shaping.systems
Some vibe coding + old/new hobby (HA, IOT, microcontrollers) + eInk + my domain expertise... Claude claims to have had fun as well: ;) "This has been a genuinely fun build— from "blank panel / SCL held low" all the way to a blue-noise-dithered, OKLab-tuned, area-normalized colour weather station.🌤️"
Do you like shaders? Do you hate shaders? Do you use shaders frequently? Khronos wants your tasty shader feedback. www.surveymonkey.com/r/LKWFQ3M
2026 Shader Ecosystem Survey
Take this survey powered by surveymonkey.com. Create your own surveys for free.
surveymonkey.com
Funny coincidence: reading about all those "cyberdecks" (super cool gadgets, and who doesn't love gadgets?) I decided to "build" (=configure) one for myself, also as a laptop replacement, from... the first Steam Deck I had sitting idle in my closet since purchase. 😅 SD card boot to Debian, 1/
...and they said their brain is incapable of operating this way. They truly put effort and felt "guilty" about it. I 100% believe them. Our brains are wired differently, sometimes at a fundamental level. We think, conceptualize, visualize, and operate differently. Hundreds of different axes. 3/3
tech artists, who implemented complex unrolled parallax occlusion mapping + mip-based blur for "cracked ice shader"... all in noodle graphs! Terrible, unreadable, slow mess, they were aware of it, asking me to port it to code. They attempted multiple times to switch to code, with our help... 2/3
I agree with all your points, and hate noodle graphs when I "have to" use them myself (like in audio environments). But IMO you are committing a common tech fallacy: missing the "human element". Some people have (irrational?) resistance against text/code. I worked with incredibly technical 1/3
I already had 2 people contact me (email, linkedin) saying that a model (I think both were ChatGPT?) urged them to contact me with some technical questions/suggestions as I am an expert on the topic. Hallucinations are funny, though. And Gemini emphasizes my work at Google. :D
austin@aparker.io · 2mo agofun? I am in there. intheweights.com
