#openvdb

[govindhtech]

NVIDIA fVDB: AI-ready Virtual-world Deep-learning Framework

At SIGGRAPH, NVIDIA unveiled fVDB, a new deep-learning framework designed to produce virtual worlds that are AI-ready.

fVDB

It is an open-source deep learning framework for sparse, large-scale, high-performance spatial intelligence that was created by NVIDIA. On top of OpenVDB, it constructs NVIDIA-accelerated  AI operators to enable 3D generative AI, neural radiance fields, digital twins at the scale of reality, and other features.

The fVDB PyTorch extension can be accessed through the fVDB Early Access program.

Generative Artificial Intelligence with Spatial Perception

It offers high resolution and large datasets for 3D deep learning infrastructure. Based on the VDB format, it integrates essential AI operators into a unified, cohesive framework. The foundation for generative physical  AI with spatial intelligence is provided by fVDB.

Superior Capabilities, Elevated Resolution

On top of NanoVDB, which offers OpenVDB GPU acceleration, are created fVDB AI operators. Real-time optimised ray tracing and sparse convolution are among the functions supported by the framework. Throughput of data processing is maximised and memory footprint is minimised with fVDB, enabling faster training and real-time inference.

Smooth Integration

It has the ability to read and write VDB datasets right out of the box if you’re already using the VDB format. It is compatible with various tools and libraries, including the Kaolin Library for 3D deep learning and Warp for Pythonic spatial computation. It’s easy to integrate fVDB into your current  AI workflow.

Examine Important Elements

Entire Operator Set

Differentiable operators such as convolution, pooling, attention, and meshing are available in fVDB and are specifically tailored for high-performance 3D deep learning applications. With the help of these operators, you may create intricate neural networks for spatial intelligence applications, such as 3D generative modelling and large-scale point cloud reconstruction.

Quicker Ray Tracing

To provide quick and precise ray tracing, it makes use of the Hierarchical Digital Differential Analyser (HDDA) technique included in OpenVDB. Neural radiance fields (NeRFs) can be quickly trained at the city-scale to produce ray-traced visualisations.

Sparse Convolution Optimisation

Sparse convolution operators in fVDB are capable of handling large 3D datasets. For activities like volumetric data analysis and physics simulation, fVDB provides quick and high-accuracy spatial data processing by optimising memory access patterns and computing burden.

Coming Soon: fVDB NIMs

Soon, NVIDIA NIM inference microservices with fVDB capability will be available, allowing developers to integrate the fVDB core architecture into USD processes. In NVIDIA Omniverse, fVDB NIMs produce geometry based on OpenUSD.

It is built on top of OpenVDB, the industry standard framework for modelling and visualising sparse volumetric data, like smoke, water, fire, and clouds.

Physical AI that is generated through generative means, like self-driving cars and real-world robots, must possess “spatial intelligence,” or the capacity to perceive, comprehend, and act in three dimensions.

It is crucial to capture both the vast scope and incredibly minute details of the environment we live in. However, it is challenging to turn reality into a virtual representation for  AI training.

There are numerous ways to gather raw data for real-world settings, including lidar and neural radiance fields (NeRFs). This data is translated into huge, AI-ready settings that are displayed in real time by fVDB.

The debut of fVDB at SIGGRAPH marks a significant advancement in the ways that sectors can profit from digital twins of the real world, building on a decade of progress in the OpenVDB standard.

Agents are trained in virtual worlds that are realistically scaled. Drones are used to collect city-scale 3D models for disaster preparedness and climate science. These days, smart cities and metropolitan areas are even planned using 3D generative  AI.

By utilising this, industries may leverage spatial intelligence at a greater scale and resolution than previously possible, resulting in even more intelligent physical AI.

Based on NanoVDB, a GPU-accelerated data structure for effective 3D simulations, the framework constructs NVIDIA-accelerated  AI operators. Convolution, pooling, attention, and meshing are some of the operators in this group; they are all intended for use in high-performance 3D deep learning applications.

Businesses can create sophisticated neural networks for spatial intelligence, such as 3D generative modelling and large-scale point cloud reconstruction, by using AI operators.

The outcome of a protracted endeavour by NVIDIA’s research team, It is currently utilised to assist projects under NVIDIA Research, NVIDIA DRIVE, and NVIDIA Omniverse that necessitate high-fidelity representations of expansive, intricate real-world environments.

Principal Benefits of fVDB

Larger: Four times the geographical dimension of earlier frameworks

Quicker: 3.5 times quicker than earlier frameworks

Interoperable: Companies have complete access to enormous real-world datasets. VDB datasets are read into full-sized 3D environments with fVDB. Real-time rendered and AI-ready for developing spatially intelligent physical  AI.

Greater power: Ten times as many operators as previous frameworks. By merging features that previously needed several deep-learning libraries, fVDB streamlines procedures.

It will soon be offered as microservices for NVIDIA NIM inference. Three of the microservices will let companies integrate fVDB into OpenUSD workflows and produce AI-ready OpenUSD geometry in NVIDIA Omniverse, a platform for developing generative physical  AI applications for industrial digitalisation. They are as follows:

fVDB Mesh Generation NIM: Creates virtual 3D worlds based on reality

fVDB NeRF-XL NIM: Utilising Omniverse  Cloud APIs, it creates extensive NeRFs in OpenUSD.

fVDB Physics Super-Res NIM Produces an OpenUSD-based, high-resolution physics simulation by performing super-resolution.

OpenVDB, a key technology utilised by the visual effects industry, has won many Academy Awards in the last ten years while being headquartered in the Academy Software Foundation. Since then, its applications in industry and science have expanded to include robots and industrial design in addition to entertainment.

NVIDIA keeps improving the OpenVDB library, which is available for free. The startup released NanoVDB four years ago, which gave OpenVDB GPU capability. This resulted in an order-of-magnitude speed increase, making real-time simulation and rendering possible as well as quicker performance and simpler programming.

NeuralVDB

A large-scale volume representation using AI-enabled data compression technology is called NeuralVDB. Compared to OpenVDB, the industry-standard library for modelling and rendering sparse volumetric data, like water, fire, smoke, and clouds, it offers a noticeable increase in efficiency.

With the release of NeuralVDB two years ago, NVIDIA expanded its machine learning capabilities beyond NanoVDB to compress VDB volumes’ memory footprint by up to 100 times. This has made it possible for researchers, developers, and producers to work with incredibly complicated and huge datasets.

On top of NanoVDB, fVDB creates  AI operators to enable spatial intelligence at the scale of reality. Submit an application to the fVDB PyTorch extension early-access program. Additionally, fVDB will be accessible through the OpenVDB GitHub repository.

Read more on Govindhtech.com

[my3dartblog]

Hoping I got everything sorted correctly and it all works fine - I'm still getting the hang of this.This pack is completely free, however, if you like the set and find it useful, please consider heading to my ko-fi to leave a tip! [ DAYDREAM CLOUD PROPS & SHADERS ] A set of 3 .obj props and matching volume shader presets. You'll need to load a .vdb for them to function (unless you'd prefer using another shader altogether - I haven't tried this myself, but it could have some interesting results!). The previews use smoke1.vbd from OpenVDB's which you can grab from their site for free (link in the readme.txt) [ DAYDREAMS RENDER PRESETS ] 6 hazy, grainy, and atmospheric render presets and environment lighting maps making use of the ground fog feature for a little extra. - - - Not recommended for Draw Dome without a DoF Camera, however all maps are 4000x2000px - - - For max fog effect, use a ground level camera at an upward angle [ CHEAP LOVE ]‥Bright, soft, and low on saturation with lavender-mint tones, a hint of pink, and a ground fog to match. [ MESSY HEARTS ]‥Bright with pink and violet hues, this environment retains a softer look while keeping a somewhat higher contrast level. [ OUTCAST ]‥A saturated, blue-intense environment with a slight color-bleed and a level of grain varying by texture use. [ HANGDOG ]‥Hazy and saturated sepia with a pink ground fog. [ BANG BANG ]‥Violet saturated with a white-point shift towards pinks for a little extra pop and a hazy pink ground fog. [ CANE SODA ]‥An extremely saturated, neon color-bleed environment. This map brings out greens and yellows in textures.Ground Fog version adds a little extra lavender to the mix. The pack also includes a readme file with setting details and a folder of quick-view renders for reference. Coming soon: https://3d-stuff.net/ #daz3d #dazstudio #3drender #3dart #daz3dstudio #irayrender #3dartwork #blender #blenderrender #blenderart #noaiart #noaiwriting #noai https://3d-stuff.net/

[ultimatepad]

“Cloud Islands”

Matej Hosek Architect

[irayrender]

Tornado data from a weather simulation done on BlueWaters Supercomputer at National Center for Supercomputing Applications, NCSA, rendered in iray + IndeX Direct using the new iray 2021.0.1 release, also available in Omniverse of course. (Retrospect: The terabyte simulation sequence was already used in a real-time distributed data rendering demo at GTC 2016 using IndeX in Paraview: https://blogs.nvidia.com/blog/2016/06/19/index-paraview/ . Newer data hopefully coming up soon, too)

[martinolsson]

[fmahalem]

[thecornellbox]

"Migration" inspired by original painting from @juliedillonart (https://www.deviantart.com/juliedillon/art/Migration-145047061)

[dustmite2013]

A little random experiment with openVDB. #typography #graphicdesign #playful #clouds #holysmoke #3D #wip #prayer #openvdb

[algomystic]

fluid works 

[amazingarchitecturewebsite]

Cloud Islands project by Matej Hosek @hosek_matej. Read more: Link in bio! Tools used: Autodesk 3ds Max , Corona Renderer, Adobe Photoshop, KitBash3D, OpenVDB #sea #nature #island #архитектура www.amazingarchitecture.com ✔ A collection of the best contemporary architecture to inspire you. #design #architecture #amazingarchitecture #architect #arquitectura #luxury #realestate #life #cute #architettura #interiordesign #photooftheday #love #travel #construction #furniture #instagood #fashion #beautiful #archilovers #home #house ‎#amazing #picoftheday #architecturephotography ‎#معماری (at Planet Earth) https://www.instagram.com/p/CQZG5h7hONB/?utm_medium=tumblr

[irayrender]

The gigantic dwarf galaxy shows many supernovae in the center and located in the center’s surrounding/rotating disc of gas and material. These supernovae spill out filaments into space. Lights highlight the supernovae. Notice how subtle the volume is lit around the supernovae by the lights.

[martinolsson]

[rendersfarm]

Autodesk has released 3ds Max 2024

Autodesk has released 3ds Max 2024. New features include:

OCIO colour management is now available as a technology preview, bringing modern colour management solutions throughout Max. 

A new boolean modifier is available that allows users to incorporate boolean operations into the modifier stack. 

An OpenVDB meshing method is included in the new boolean modifier, providing an alternate wat to calculate operations. OpenVDb booleans are topologically independent of the source geometry and they will often yield good results even when mesh errors are present that would cause standard boolean techniques to fail. 

The array modifier is updated with a new Phyllotaxis distribution technique. This type of pattern is frequently seen in nature, particularly in plants and trees. 

The array modifier includes a new Material ID rollout that can control the materials IDs on a per-face or per-element basis. Several modes are available including random, sequenced, and the ability to change the ID of the first and last clone in an array and randomise the remainder. 

The array modifier can now use a progressive transform mode that increments the transform value for each element. 

A new controller allows the user to add multiple Transform controllers to an object as layers of blended animation data. This lets animators add unique animation data to each layer, as well as allows them to control how the layers blend together.

A new material node allows the user to switch between multiple materials using a numeric index. 

The motion paths feature that allows users to visualise animation paths has been updated so that it can now be used for all controller types and it works with the new list controllers. 

The slate material editor has been overhauled to use QT. As a result, it’s much more customisable and dockable, as well as promising significant speed improvements. 

The material editor now includes a new Compound node that allows you to create sub-graphs. These can be used to organise large graphs to aid legibility. Creating a compound is as simple as right-clicking on a selection of nodes and selecting Package Nodes in Compound, 

The Modifier list has been improved to use QT and now includes a list search feature so that users can start typing to find the modifier they need. 

The Symmetry Modifier‘s default Mirror Axis is now X to better align with artist’s workflows.

STL import is now an impressive 10,000 times faster, and the STL Check modifier is 5000% times faster.

Auto Smooth is now at least 10% faster

The Material Modifier can now be used on splines without them being converted to a mesh. It can now also retain explicit normals when applied to a mesh. 

Spline Welding is improved 

Retriangulation for Edit Poly and Editable poly is improved as follows: The edit poly modifier now uses the new retriangulation algorithm introduced in Max 2023.2;  Editable Poly objects and the Edit Poly modifier now automatically retriangulate faces when a vertex, edge, or face is adjusted in a way that causes edges or hidden faces to cross; New faces that are generated when performing capping are no longer left unsmoothed. Instead, faces are assigned to a new smoothing group; The new traingulation algorithm is now used for face splitting by insertion of edges, Slice, Cut, Bridge, Vertex extrusion, and edge extrusion.

3ds Max 2024 ships with Substance 2.4.10 and Arnold 5.6.0.1

Source: https://blog.superrendersfarm.com/news/autodesk-has-released-3ds-max-2024/

[dustmite2013]

A little random experiment with openVDB. #typography #graphicdesign #playful #clouds #holysmoke #3D #wip #prayer #openvdb