frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Open in hackernews

Show HN: Real-Time Gaussian Splatting

https://github.com/axbycc/LiveSplat
98•markisus•5h ago
LiveSplat is a system for turning RGBD camera streams into Gaussian splat scenes in real-time. The system works by passing all the RGBD frames into a feed forward neural net that outputs the current scene as Gaussian splats. These splats are then rendered in real-time. I've put together a demo video at the link above.

Comments

sreekotay•5h ago
This is realtime capture/display? Presumable (at this stage) for local viewing? Is that right?
markisus•5h ago
Yes realtime capture and display. Locality is not required. You can send the source RGBD video streams over IP and in fact I have that component working in the larger codebase that this was split off from. For that use case, you need to do some sort of compression. The RGB stream compression is a pretty solved problem, but the depth channel needs special consideration since "perceptual loss" in the depth space is not a well researched area.
patrick4urcloud•5h ago
nice
echelon•5h ago
OP, this is incredible. I worry that people might see a "glitchy 3D video" and might not understand the significance of this.

This is getting unreal. They're becoming fast and high fidelity. Once we get better editing capabilities and can shape the Gaussian fields, this will become the prevailing means of creating and distributing media.

Turning any source into something 4D volumetric that you can easily mold as clay, relight, reshape. A fully interactable and playable 4D canvas.

Imagine if the work being done with diffusion models could read and write from Gaussian fields instead of just pixels. It could look like anything: real life, Ghibli, Pixar, whatever.

I can't imagine where this tech will be in five years.

markisus•5h ago
Thanks so much! Even when I was putting together the demo video I was getting a little self-critical about the visual glitches. But I agree the tech will get better over time. I imagine we will be able to have virtual front row seats at any live event, and many other applications we haven't thought of yet.
echelon•5h ago
> I imagine we will be able to have virtual front row seats at any live event, and many other applications we haven't thought of yet.

100%. And style-transfer it into steam punk or H.R. Giger or cartoons or anime. Or dream up new fantasy worlds instantaneously. Explore them, play them, shape them like Minecraft-becomes-holodeck. With physics and tactile responses.

I'm so excited for everything happening in graphics right now.

Keep it up! You're at the forefront!

_verandaguy•5h ago
I know enough about 3D rendering to know that Gaussian splatting's one of the Big New Things in high-performance rendering, so I understand that this is a big deal -- but I can't quantify why, or how big a deal it is.

Could you or someone else wise in the ways of graphics give me a layperson's rundown of how this works, why it's considered so important, and what the technical challenges are given that an RGB+D(epth?) stream is the input?

markisus•5h ago
Gaussian Splatting allows you to create a photorealistic representation of an environment from just a collection of images. Philosophically, this is a form of geometric scene understanding from raw pixels, which has been a holy grail of computer vision since the beginning.

Usually creating a Gaussian splat representation takes a long time and uses an iterative gradient-based optimization procedure. Using RGBD helps me sidestep this optimization, as much of the geometry is already present in the depth channel and so it enables the real-time aspect of my technique.

When you say "big deal", I imagine you are also asking about business or societal implications. I can't really speak on those, but I'm open to licensing this IP to any companies which know about big business applications :)

corysama•4h ago
So, is there some amount of gradient-based optimization going on here? I see RGBD input, transmission, RGBD output. But, other than multi-camera registration, it's difficult to determine what processing took place between input and transmission. What makes this different from RGBD camera visualizations from 10 years ago?
markisus•3h ago
There is no gradient-based optimization. It's (RGBD input, Current Camera Pose) -> Neural Net -> Gaussian Splat output.

I'm not aware of other live RGBD visualizations except for direct pointcloud rendering. Compared to pointclouds, splats are better able to render textures, view-dependent effects, and occlusions.

sendfoods•5h ago
Please excuse my naive question - isn't Gaussian Splatting usually used to create 3D imagery from 2D? How does providing 3D input data make sense in this context?
ttoinou•5h ago
Well if you have the D channel you might as well benefit from it and have better output
markisus•5h ago
Yes, the normal case uses 2D input, but it can take hours to create the scene. Using the depth channel allows me to create the scene in 33 milliseconds, from scratch, every frame. You could conceptualize this as a compromise between raw pointcloud rendering and fully precomputed Gaussian splat rendering. With pointclouds, you have a lot visual artifacts due to sparsity (low texture information, seeing "through" objects"). With Gaussian splatting, you can transfer a lot more of the 2D texture information into 3D space and render occlusion and view-dependent effects better.
Retr0id•4h ago
How do the view-dependent effects get "discovered" from only a single source camera angle?
markisus•4h ago
Actually there are multiple source cameras. The neural net learns to interpolate the source camera colors based on where the virtual camera is. Under the hood it's hard to say exactly what's going on in the mind of the neural net, but I think it's something like "If I'm closer to camera A, take most of the color from camera A."
ttoinou•4h ago
So we’re not sure how it works exactly ?
markisus•4h ago
Yup, this is the case for all neural nets.
jayd16•3h ago
Splatting is about building a scene that supports synthetic view angles.

The depth is helpful to properly handle the parallaxing of the scene as the view angle changes. The system should then ideally "in-paint" the areas that are occluded from the input.

You can either guess the input depth from matching multiple RGB inputs or just use depth inputs along with RGB inputs if you have them. It's not fundamental to the process of building the splats either way.

yuchi•5h ago
The output looks terribly similar to what sci-fi movies envisioned as 3D reconstruction of scenes. It is absolutely awesome. Now, if we could project them in 3D… :)
tough•1h ago
Apple Vision maybe?
mandeepj•5h ago
Another implementation of splat https://github.com/NVlabs/InstantSplat
jasonjmcghee•4h ago
The quality is better, no doubt, but this method (from the paper) takes on the order of 10-45s depending on input from their table. Which is much better than 10 minutes etc.

That being said, afaict OP's method is 1000x faster, at 33ms.

markisus•4h ago
Note that the method you linked is "Splatting in Seconds" where as real-time requires splatting in tens of milliseconds.

I'm also following this work https://guanjunwu.github.io/4dgs/ which produces temporal Gaussian splats but takes at least half an hour to learn the scene.

metalrain•4h ago
How did you train this? I'm thinking there isn't reference output for live video frame to splats so supervised learning doesn't work.

Is there some temporal accumulation?

markisus•4h ago
There is no temporal accumulation, but I think that's the next logical step.

Supervised learning actually does work. Suppose you have four cameras. You input the three of them into the net and use the fourth as the ground truth. The live video aspect just emerges from re-running the neural net every frame.

corysama•4h ago
So, I see livesplat_realsense.py imports livesplat. Where’s livesplat?
IshKebab•4h ago
The README says it's closed source.
markisus•4h ago
I've tried to make it clear in the link that the actual application is closed source. I'm distributing it as a .whl full of binaries (see the installation instructions).

I've considered publishing the source but the source code is is dependent on some proprietary utility libraries from my bigger project and it's hard to fully disentangle it and I'm not sure if this project has some business applications but I'd like to keep that door open at this time.

armchairhacker•4h ago
Gaussian Splatting looks pretty and realistic in a way unlike any other 3D render, except UE5 and some hyper-realistic not-realtime renders.

I wonder if one can go the opposite route and use gaussian splatting or (more likely) some other method to generate 3D/4D scenes from cartoons. Cartoons are famously hard to emulate in 3D even entirely manually; like with traditional realistic renders (polygons, shaders, lighting, post-processing) vs gaussian splats, maybe we need a fundamentally different approach.

spyder•4h ago
Correct me if I'm wrong but looking at the video this just looks like a 3D point cloud using equal-sized "gaussians" (soft spheres) for each pixel, that's why it looks still pixelated especially at the edges. Even when it's low resolution the real gaussian splatting artifacts look different with spikes an soft blobs at the lower resolution parts. So this is not really doing the same as a real gaussian splatting of combining different sized view-dependent elliptic gaussians splats to reconstruct the scene and also this doesn't seem to reproduce the radiance field as the real gaussian splatting does.
markisus•4h ago
I had to make a lot of concessions to make this work in real-time. There is no way that I know to replicate the fidelity of "actual" Gaussian splatting training process within the 33ms frame budget.

However, I have not baked in the size or orientation into the system. Those are "chosen" by the neural net based on the input RGBD frames. The view dependent effects are also "chosen" by the neural net, but not through an explicit radiance field. If you run the application and zoom in, you will be able to see the splats of different sizes pointing in different directions. The system as limited ability to re-adjust the positions and sizes due to the compute budget leading to the pixelated effect.

whywhywhywhy•4h ago
Would be good to see how it's different from just the depth channel applied to the Z of the RGB pixels. Because it looks very similar to that.
markisus•4h ago
The application has this feature and lets you switch back and forth. What you are talking about is the standard pointcloud rendering algorithm. I have an older video where I display the corresponding pointcloud [1] in a small picture in picture frame so you can compare.

I actually started with pointclouds for my VR teleoperation system but I hated how ugly it looked. You end up seeing through objects and objects becoming unparseable if you get too close. Textures present in the RGB frame also become very hard to make out because everything becomes "pointilized". In the linked video you can make out the wood grain direction in the splat rendering, but not in the pointcloud rendering.

[1] https://youtu.be/-u-e8YTt8R8?si=qBjYlvdOsUwAl5_r&t=14

badmonster•1h ago
What is the expected frame rate and latency when running on a typical setup with one Realsense camera and an RTX 3060?
markisus•43m ago
I don't have a 3060 at hand so I'm not sure. Ideally someone with that setup will try it out and report back. There is no noticeable latency when comparing visually with standard pointcloud rendering.

With framerate, there are two different frame rates that are important. One is the splat construction framerate, which the speed that an entirely new set of Gaussian's can be constructed. LiveSplat can usually maintain 30fps in this case.

The second important splat rendering framerate. In VR this is important to prevent motion sickness. Even if you have a static set of splats, you need the rendering to react to the user's minor head movements at around 90fps for the best in-headset experience.

All these figures are on my setup with a 4090 but I have gotten close results with a 3080 (maybe 70fps splat rendering instead of 90fps).

Workers are saying 'no' to toxic environments

https://apnews.com/article/what-to-do-toxic-workplace-job-7c96353516e7b3968e21b0d7099fd877
1•mooreds•1m ago•0 comments

Virginia Woolf? Snob Richard Wright? Sexist Dostoyevsky? Anti-Semite (2019)

https://www.nytimes.com/2019/01/08/books/review/edith-wharton-house-of-mirth-anti-semitism.html
1•Tomte•3m ago•0 comments

Get ready to see a lot more mini-trucks on Colorado's roads

https://www.cpr.org/2025/05/13/law-allows-japanese-mini-trucks-colorado-roads/
1•mooreds•3m ago•0 comments

Unicode in Five Minutes ⌚ (2013)

https://richardjharris.github.io/unicode-in-five-minutes/
1•Tomte•3m ago•0 comments

Show HN: Wavesly – Check Sea and Ocean Temperature

https://www.wavesly.io
1•coolpool•3m ago•0 comments

Show HN: Smol vsc*de plugin to find in all open files via native search

https://github.com/f-person/vscode-search-in-open-files
1•fperson•4m ago•0 comments

Scott Forstall on the Future of Siri and AI (2017) [video]

https://www.youtube.com/watch?v=IiuVggWNqSA
1•bangonkeyboard•8m ago•1 comments

October the First Is Too Late

https://gwern.net/fiction/october
3•adiabatty•8m ago•0 comments

Catalogue of Novel Operating Systems

https://github.com/prathyvsh/os-catalog
1•lioeters•9m ago•0 comments

Keyboard Sounds – Add sound effects to your typing experience

https://keyboardsounds.net/
2•nfiscaletti•12m ago•0 comments

Rafale maker Dassault shares plummet, Avic Chengdu maker of J-10 stock soar

https://www.hindustantimes.com/business/rafale-maker-dassault-aviation-shares-plummet-amid-india-pakistan-tension-chinas-cac-stock-soars-101747108446631.html
2•teleforce•12m ago•0 comments

My full statement regarding DOOM Eternal (2022)

https://medium.com/@mickgordon/my-full-statement-regarding-doom-eternal-5f98266b27ce
5•microsoftedging•16m ago•0 comments

LLM functions in TypeScript: a composable pattern for prompt/LLM/parse/execute

https://medium.com/llm-exe
2•llm-exe•18m ago•1 comments

The Dumbest Move in Tech: Laying Off Developers Because of AI

https://ppaolo.substack.com/p/the-dumbest-move-in-tech-right-now
2•paolop•21m ago•3 comments

Gitlab 18.0 Released with Gitlab Duo for Premium and Ultimate

https://about.gitlab.com/releases/2025/05/15/gitlab-18-0-released/
2•bmaupin•21m ago•0 comments

Keep calm (but delete your nudes): the new rules for travelling to America

https://www.theguardian.com/us-news/2025/may/15/travel-trump-america-us-border-detentions
2•n1b0m•23m ago•0 comments

Has Grok lost its mind and mind-melded with its owner?

https://garymarcus.substack.com/p/has-grok-lost-its-mind-and-mind-melded
5•NotInOurNames•23m ago•0 comments

The current state of TLA⁺ development

https://ahelwer.ca/post/2025-05-15-tla-dev-status/
4•todsacerdoti•23m ago•0 comments

OpenAI to Z Challenge

https://openai.com/openai-to-z-challenge/
1•Develra•24m ago•0 comments

Why untested AI-generated code is a crisis waiting to happen

https://leaddev.com/software-quality/why-untested-ai-generated-code-is-a-crisis-waiting-to-happen
1•mooreds•25m ago•0 comments

Ask HN: Is Substack the New Medium?

6•behnamoh•26m ago•1 comments

What's New in the Oracle Solaris 11.4.81 CBE Release

https://blogs.oracle.com/solaris/post/whats-new-in-the-oracle-solaris-11481-cbe-release
1•naves•26m ago•0 comments

About Asteroids, Atari's biggest arcade hit

https://www.goto10retro.com/p/about-asteroids-ataris-biggest-arcade
1•rbanffy•27m ago•0 comments

China launches satellites to build the first supercomputer in orbit

https://www.scmp.com/news/china/science/article/3310506/china-launches-satellites-start-building-worlds-first-supercomputer-orbit
4•giuliomagnifico•27m ago•0 comments

Open-sourcing Pyrefly: A faster Python type checker written in Rust

https://engineering.fb.com/2025/05/15/developer-tools/open-sourcing-pyrefly-a-faster-python-type-checker-written-in-rust/
1•vquemener•28m ago•0 comments

Smart Money Loses to Retail Crowd That Bet on Epic Stock Rebound

https://www.bloomberg.com/news/articles/2025-05-15/smart-money-loses-to-retail-crowd-that-bet-on-epic-stock-rebound
2•littlexsparkee•28m ago•1 comments

Windsurf SWE-1: Our First Frontier Models

https://windsurf.com/blog/windsurf-wave-9-swe-1
10•arittr•29m ago•0 comments

Show HN: Full LLM fine-tuning and evaluation guide

https://ubiai.gitbook.io/llm-guide
1•Mesterniz•29m ago•0 comments

Dr. Dobb's Journal Interviews Jef Raskin (1986)

https://computeradsfromthepast.substack.com/p/dr-dobbs-journal-interviews-jef-raskin
4•rbanffy•33m ago•0 comments

Renewable power reversing China's emissions growth

https://arstechnica.com/science/2025/05/analysis-shows-that-chinas-emissions-are-dropping-due-to-renewables/
4•moelf•35m ago•1 comments