/Tech11h ago

Janusch Patas showcases LichtFeld Studio rendering 260 million Gaussians at 33 fps using a new Level-of-Detail pipeline

Converting PLY files to RAD takes only five minutes.

12162237010.9K
MrNeRF
Dmytro Mishkin 🇺🇦
martin_casado#480
Original postDmytro Mishkin 🇺🇦#1645
MrNeRF@janusch_patas

Everybody is stepping into the LOD game, and so is LichtFeld Studio.

Preview of 260M Gaussians streaming into the viewer live.

I use the RAD format which is processed on GPU within LichtFeld. It can be also simply dumped straight into the spark.js web viewer, albeit it will die at that amount of Gaussians.

What other solutions don't tell you is that they need hours to preprocess a 3DGS ply to make it streamable. This was just a ply exported to RAD by LichtFeld Studio's convert tool (took 5 min at that size) and it is immediately ready to stream.

In the comments there is a smaller dataset with 103M Gaussians that streams on startup into the viewer. Both datasets were created by Andrii Shramko.

Let's see how far I can push this (need bigger datasets)

7:14 AM · Jun 10, 2026 · 10.9K Views
Sentiment

Positive users praise LichtFeld Studio's live streaming of 260M Gaussians with Fast RAD LOD as technically cool and innovative, while negative users sarcastically dismiss it as inferior to other tools.

Pos
50.0%
Neg
50.0%
2 comments with sentiment.
Cluster Engagement
Posts from X
Most Activity
Most Activity
VIEWS674LIKES10
MrNeRF@janusch_patas

Here it shows the loading into the viewport with 103m Gaussians.

11hViews 674Likes 10
BOOKMARKS1
Lodewijk Luijt@LodewijkLuijt

@janusch_patas This dataset is pretty big.... https://huggingface.co/datasets/wildflow/sweet-corals

10hViews 18Bookmarks 1
RETWEETS12
MrNeRF@janusch_patas

Everybody is stepping into the LOD game, and so is LichtFeld Studio.

Preview of 260M Gaussians streaming into the viewer live.

I use the RAD format which is processed on GPU within LichtFeld. It can be also simply dumped straight into the spark.js web viewer, albeit it will die at that amount of Gaussians.

What other solutions don't tell you is that they need hours to preprocess a 3DGS ply to make it streamable. This was just a ply exported to RAD by LichtFeld Studio's convert tool (took 5 min at that size) and it is immediately ready to stream.

In the comments there is a smaller dataset with 103M Gaussians that streams on startup into the viewer. Both datasets were created by Andrii Shramko.

Let's see how far I can push this (need bigger datasets)

11hViews 10.9KLikes 162Bookmarks 70
REPLIES1
Rafael Spring@Rafael_L_Spring

@janusch_patas Very cool!! Are the LODs computed at optimization time or can any splat get a LOD hierarchy and streamability?

10hViews 102Likes 2
Mykhailo Moroz@Michael_Moroz_

@janusch_patas Ah right its the Lublin splat from Shramko, I already have it downloaded😅, nevermind

10hViews 38Likes 1
Mykhailo Moroz@Michael_Moroz_

@janusch_patas Is that 260M splat open by the way? I'd love to mess around with it too

10hViews 116
MrNeRF@janusch_patas

@Rafael_L_Spring any splat. See the bhatt algo from spark.js

10hViews 78Likes 2
Irfan Kahvedzic@ir_foo

@janusch_patas Speed of light being the ultimate speed limit is probably how the universe does LOD lol

10hViews 12
MrNeRF@janusch_patas

@Michael_Moroz_ Yeah, sorry for the late response :) That's the dataset.

10hViews 3Likes 1
Rob Sloan@robmakes3d

@janusch_patas ...if only it were as good as other tools /sarcasm

9hViews 5