Sergiu Zboras asked how USD compression compares to .sog and .splat on my last post. I was curious, so I ran the test. Same 3DGS capture, four formats: .lcc 541MB → .ply 749MB → .usd 712MB → .sog 41MB USD sits close to PLY. .sog wins by a country mile for streaming and distribution. The value of USD is composability though. A layerable, referenceable, variant-capable scene object that lives natively in the same pipeline as every other piece of geometry in a professional production workflow. That's a different use case entirely. One honest caveat: I can't render it visually yet. The ParticleField3DGaussianSplat schema landed in OpenUSD 26.03 in March 2026 and the renderers haven't caught up. The pipeline works. The viewers are following. #RealityCapture #OpenUSD #SpatialData #GaussianSplatting #3DCapture
More Relevant Posts
-
💥 HY-World 2.0 just dropped ♠ and it turns any photo or casual video into real, editable 3D worlds 🔹 One image or phone video → full Gaussian Splats + meshes 🔹 Fully navigable, importable into Unity/Unreal/Blender 🔹 WorldMirror 2.0 does instant reconstruction in a single pass 🔹 First open-source model that actually beats closed-source rivals 🔹 I just tested it live with my own image. 🔹 The result is insane. 🔥 Watch the full demo below https://lnkd.in/gmkC2XrN
To view or add a comment, sign in
-
🎥𝗚𝗮𝘂𝘀𝘀𝗶𝗮𝗻 𝗦𝗽𝗹𝗮𝘁𝘁𝗶𝗻𝗴 𝗶𝘀 𝗻𝗼 𝗹𝗼𝗻𝗴𝗲𝗿 𝗷𝘂𝘀𝘁 𝗮𝗻 𝗥&𝗗 𝘁𝗼𝗽𝗶𝗰. 𝗜𝘁’𝘀 𝗮𝗰𝘁𝗶𝘃𝗲𝗹𝘆 𝗿𝗲𝘀𝗵𝗮𝗽𝗶𝗻𝗴 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗽𝗶𝗽𝗲𝗹𝗶𝗻𝗲𝘀 𝘁𝗼𝗱𝗮𝘆. At NAB Show we brought together early adopters to break down how 3D Gaussian Splatting (3DGS) is moving from experimentation 𝗶𝗻𝘁𝗼 𝗿𝗲𝗮𝗹-𝘄𝗼𝗿𝗹𝗱 𝘂𝘀𝗲 across Virtual Production, VFX, and beyond. In this panel, Fernando Rivas-Manzaneque, Joao Krambeck from RCS Studios, Kathryn Brillhart, and Michael Rubloff from Radiance Fields (Gaussian Splatting and NeRFs), share 𝗵𝗼𝘄 𝘁𝗲𝗮𝗺𝘀 𝗮𝗿𝗲 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝘂𝘀𝗶𝗻𝗴 𝟯𝗗𝗚𝗦 𝘁𝗼𝗱𝗮𝘆: ➡️What Gaussian Splats really are (and why they’re different from traditional 3D capture) ➡️How teams are using them in previs, ICVFX, 𝗮𝗻𝗱 𝗽𝗼𝘀𝘁-𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀 ➡️Real-world 𝗲𝘅𝗮𝗺𝗽𝗹𝗲𝘀 𝗼𝗳 𝗰𝗮𝗽𝘁𝘂𝗿𝗶𝗻𝗴 𝗹𝗼𝗰𝗮𝘁𝗶𝗼𝗻𝘀 and building photoreal digital environments ➡️Where 𝘁𝗵𝗲 𝗯𝗶𝗴𝗴𝗲𝘀𝘁 𝘁𝗶𝗺𝗲 𝗮𝗻𝗱 𝗰𝗼𝘀𝘁 𝘀𝗮𝘃𝗶𝗻𝗴𝘀 are showing up in production ➡️The practical challenges, from 𝗱𝗮𝘁𝗮 𝗵𝗮𝗻𝗱𝗹𝗶𝗻𝗴 𝘁𝗼 𝗼𝗻-𝘀𝗲𝘁 𝗶𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻. If you’re working in Virtual Production, this is where things start to get very real: faster environment capture, fewer bottlenecks, and new ways to bridge physical and digital worlds.🙌 👀𝗪𝗮𝘁𝗰𝗵 𝘁𝗵𝗲 𝗳𝘂𝗹𝗹 𝗽𝗮𝗻𝗲𝗹 𝗵𝗲𝗿𝗲: https://lnkd.in/eMQcZAfB #VirtualProduction #GaussianSplatting #3DGS #VFX #ICVFX #FilmmakingTechnology #DigitalEnvironments #ProductionPipeline #NABShow
To view or add a comment, sign in
-
We’re excited to welcome Roc Lee (he/him), Senior Sound Designer at Clean Cuts Interactive , to speak about building custom workflow tools in FMOD! Based in Washington, D.C., Roc specializes in creating audio systems that support gameplay and storytelling, with credits including Asgard’s Wrath 2, Saber Punks, and Sword Reverie. He’s also spoken at Game Developers of Color Expo, GameSoundCon, and Audio Engineering Society AVAR, while helping organize the DMV Game Audio meetup community. Talk: Fix Your Own Workflow: Building the FMOD Tools You Wish Existed From repetitive FMOD tasks to daily-use custom tools, Roc will share how small frustrations led him to create scripts including: - Event Renamer - Bank Builder - Memory Location Toolbar - Import Tool Whether you’re new to scripting or deep into technical audio, this talk is a reminder that your workflow frustrations are worth solving. No scripting experience needed! Link to event available on the QR code #asiangameaudionetwork #fmod #workflow #gamedev #gameaudio
To view or add a comment, sign in
-
-
𝗝𝗼𝗶𝗻 𝘂𝘀 𝗹𝗶𝘃𝗲 𝘁𝗼 𝗱𝗶𝘀𝗰𝗼𝘃𝗲𝗿 𝗵𝗼𝘄 𝟯𝗗 𝗚𝗮𝘂𝘀𝘀𝗶𝗮𝗻 𝗦𝗽𝗹𝗮𝘁𝘁𝗶𝗻𝗴 𝗶𝘀 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗯𝗲𝗶𝗻𝗴 𝘂𝘀𝗲𝗱 𝗶𝗻 𝗿𝗲𝗮𝗹 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀 🚀 As part of our ongoing '𝗗𝗶𝘀𝗰𝗼𝘃𝗲𝗿 𝟯𝗗𝗚𝗦 𝘄𝗶𝘁𝗵 𝗩𝗼𝗹𝗶𝗻𝗴𝗮' live sessions, the Volinga team regularly walks through real-world Gaussian Splatting workflows, from capture and scene setup to Unreal Engine integration and on-set deployment. 𝗢𝘂𝗿 𝗻𝗲𝘅𝘁 𝘀𝗲𝘀𝘀𝗶𝗼𝗻 𝘁𝗮𝗸𝗲𝘀 𝗽𝗹𝗮𝗰𝗲 𝗻𝗲𝘅𝘁 𝗧𝘂𝗲𝘀𝗱𝗮𝘆, 𝗠𝗮𝘆 𝟭𝟵𝘁𝗵, and it’s a great opportunity to explore these workflows through practical demos, real use cases, and open discussion with the team. 𝗪𝗲’𝗹𝗹 𝗰𝗼𝘃𝗲𝗿: ☑️Real production workflows using 3D Gaussian Splatting ☑️Step-by-step guidance from capture to Unreal Engine ☑️Plugin setup, sample projects, and common pitfalls ☑️Live Q&A with the Volinga team Whether you’re working in virtual production, VFX, broadcast, or real-time environments, this session is designed to help you understand how these workflows operate in practice, not just in theory. 👉 𝗥𝗲𝗴𝗶𝘀𝘁𝗲𝗿 𝗵𝗲𝗿𝗲: https://lnkd.in/e5BqnB-P #Volinga #3DGS #GaussianSplatting #VirtualProduction #Webinar
To view or add a comment, sign in
-
🔵 My team needed to simulate +50,000 dynamic particles for a mapping show... So I made my own tool. I created this real-time modular particle collision system running fully in the browser using WebGPU. The goal was to add as many parameters as possible in the GUI to let the visual artists have complete control over how the particles behave. Using those inputs, we can easily change the shape of a particle, its speed, or its color. I also added a way to create collision shapes and an effect to repulse particles. Hope you liked that :) #interactive #projection #mapping #pixijs #webgpu #realtime #art #creativecodding
To view or add a comment, sign in
-
We ran MForja on cinematic footage. The result stopped me. 9.5 MB source. 647 KB output. 93.2% file size reduction. VMAF 99.99. The clip is a low-light scene, a man putting a hat on a horse under soft candlelight, from Film Supply. That kind of footage is compression's worst case: dark gradients, organic textures, film grain, subtle motion throughout. The entropy shim inserted upstream of libx265. No codec swapped. No hardware touched. No changes to any downstream infrastructure. VMAF 99.99 means the output is, for all measurable purposes, indistinguishable from the source. Here is what 93.2% reduction means at scale. If you are a platform storing 1 petabyte of video, that is roughly 930 terabytes back. If you are a drone manufacturer shipping 1,000 units a year, that is a different conversation with every customer who asks about onboard storage or transmission bandwidth. The four pillars of signal waste, storage, bandwidth, power, codec efficiency, all move together when the entropy is right at the source. This is what Video Distillation looks like in practice. mforja.com
To view or add a comment, sign in
-
Quick one for the Blackmagic Design DaVinci Resolve crowd: how are you getting your diffusion right now? Some folks use bloom nodes, glow effects, or just softly roll off the highlights. We felt there wasn't a straightforward method, so we built a tool called DigiDiff that handles it all in one place. The diffusion is physically accurate, so it behaves the way physical diffusion does. And you get a lot more control without the messy workflow! Worth a look if you're in the Blackmagic Design world. Join us on https://lnkd.in/gDss_Zqz to see how creators are using DigiDiff—and share your own results.
To view or add a comment, sign in
-
Distortion Grids Database is a structured library built for real post-production workflows. It includes: - distortion grids - ready-made lens presets - LensDistortion nodes for Nuke - 3DEqualizer project files Upload assets → receive ratings, discounts, and a reason to keep building value for the whole community. Our goal is to turn lens data from a one-time production cost into a reusable production resource — making matchmove and VFX workflows more reliable, faster, and easier to run. Enjoy! And Have a Good Solve! #Matchmove #3DTracking #CameraTracking #VFX #CGI #VisualEffects #FilmProduction #MatchmoveArtist #MotionTracking #VFXCommunity
Distortion Grids Database
To view or add a comment, sign in
-
Manually placing reverb zones is a thing of the past. I built Fibonacci Acoustic Scanner Pro — a runtime acoustic plugin for Unity + Wwise. Zero baked data. Pure physics-based simulation. What you see in this video: → Occlusion calculated in real time as the source moves behind walls → Diffraction path bending around corners — live as you peek out → Reverb Amount driven by actual geometry, not trigger zones → Source picked up — occlusion clears instantly, reverb keeps updating No zones. No triggers. Just raycasts + Wwise RTPC. Free to download — pay what you want. 🔗 itch.io: https://lnkd.in/gBd4XHd3 🔗 GitHub: https://lnkd.in/gW_M8r9T #gameaudio #unity3d #wwise #sounddesign #gamedev #plugin #indiedev
To view or add a comment, sign in
-
Well designed! This is just the start 😉 To 3D animators and indie game devs, you’ll learn how easy and accurate the mocap function is when you try it!
I barely post on LinkedIn, but I've got something worth sharing. Spent the last month cooking with the team on PINOC. It's live today. One video in, motion out as FBX / GLB. It also generates custom Gaussian splat model so you can preview the motion with any character you want. No guessing what you're going to get. Built for the folks priced out of mocap studios, burnt out from hand-keyframing every gesture, or stuck searching Mixamo for a motion that just isn't there. Free for all during the beta test. This is v1.0. More on the way. Try PINOC today at: https://viggle.ai/pinoc
To view or add a comment, sign in
Thanks for providing the info. Very interesting. SOG indeed is crazy good at compressing so much data.