Epic Games has announced the release of Unreal Engine 5.3. It can be downloaded from the company’s website.
This release includes a wide range of improvements, expanding the capabilities and possibilities of Unreal Engine ver. 5 for game developers and creators across the industry. In addition to core rendering, developer iterations, and virtual production toolsets, we also introduce experimental new features for rendering, animation, and simulation, allowing creators to extend creative capabilities inside UE5 without having to go back and forth to external applications.
New features introduced in Unreal Engine 5.3
Improved core rendering features in Unreal Engine 5
This release continues to refine all of UE5’s core rendering capabilities, with the continued goal of enabling developers to take advantage of them at high quality in games running at 60fps on next-gen consoles.
Specifically, Nanite has improved performance for masked materials such as foliage, and a new “Explicit Tangents” option allows users to represent a wider range of surfaces. Also, with hardware ray tracing, Lumen now has enhanced features such as multiple reflection bounces and faster performance on consoles.
Other notable advances include production-ready virtual shadow maps (VSM), temporal super resolution (TSR), hair grooms, path tracing, and Substrate.
Multi process cook
Other useful improvements include the ability to take advantage of additional CPU and memory resources when converting content from internal UE formats to platform-specific formats. Therefore, it can significantly reduce the time it takes to get cooked output from a build farm server or local workstation.
Enabling multi-process cooking allows creators to start a subprocess that performs the cooking part in parallel with the main process. Developers can choose the number of subprocesses to run on a single machine.
Cine Camry Rail
Filmmakers can now emulate the workflow and results of traditional camera movement along a track or dolly thanks to the new Cine Camry Rail Actor.
The new Cine Cam Rig Rail Actor allows for more sophisticated control than the existing Rig Rail with various control points along the path, including the ability to choreograph settings such as camera rotation, focal length, and focus distance. . Both in-editor and VCam workflows are supported.
VCam enhancements
The VCam system has received many enhancements. The ability to speed up iterations by reviewing takes directly on the iPad, and the ability to facilitate collaborative VCam shooting by simultaneously streaming different VCam outputs to different team members (for example, the camera (with operator camera control), and the ability to easily capture fast-moving action by recording at a slower frame rate and playing back at normal speed.
Experimental features
In addition to these updates to the core toolset, Unreal Engine 5.3 also introduces some great new experimental features. We plan to further develop these features in future releases. These features are not yet recommended for use in virtual production.
Cinema-quality volumetric rendering
Two new features, Sparse Volume Texturing (SVT) and Heterogeneous Volume Path Tracing, introduce many new features for volumetric effects such as smoke and fire.
Sparse volume textures store baked simulation data representing volumetric media. Users can also import and simulate OpenVDB (.vdb) files created with Niagara or other 3D applications.
Additionally, support for more complete volume rendering is now available as an experimental feature in Path Tracer. This provides high-quality volumetric rendering possibilities such as global illumination, shadows, and scattering for cinematics, movies, episodic television, and other linear content production created directly in UE5.
For real-time use cases such as gaming and virtual production, experimental functionality can also be initiated with SVT for the reproduction of volumetric elements. However, performance is currently limited and highly content dependent.
Orthographic rendering
Starting with UE 5.3, there are plans to introduce orthographic rendering. It is useful for visualizing architectural and manufacturing projects, as well as offering orthographic projection as a stylistic camera option for games.
The new update adds parity between perspective and orthography focuses attention on multiple areas of the engine. Most of UE5’s latest features are expected to be supported, including Lumen, Nanite, shadows, and Temporal Super Resolution (TSR). Orthographic rendering is also available in the Unreal Editor, allowing users to make updates in live settings.
Skeletal editor
The new Skeletal Editor allows animators to use a variety of tools when working with Skeletal Meshes, including tools to paint skin weights. So whether it’s a quick prototype or final rigging, more character workflows can be performed entirely within Unreal Editor without having to switch back and forth to the DCC application. This means working without losing context and iterate faster.
Panel-based Chaos cloth with ML simulation
In addition to being designed to add more creative workflows directly into UE, this release also includes several updates to Chaos Cloth. Introducing a new panel cloth editor and skin weight transfer algorithm, and adding XPBD (Extended Position-Based Dynamics) constraints as the basis for future cloth generation in the engine. This enables a non-destructive cross-simulation workflow that allows designers and makers to trade off speed and accuracy. Furthermore, the panel-based cloth allows for better-looking simulations.
By using the new Panel Cloth Editor and ML Deformer Editor, creators can now simulate cloth and cache it in the engine.
nDisplay support for SMPTE ST 2110
This update added experimental support for nDisplay in SMPTE ST 2110 using NVIDIA hardware. Also, support for Rivermax SDK was added. This is to prepare for the next generation of LED production stages. This will serve as the basis for a variety of hardware configurations that will open up new possibilities for LED stages. The superior configuration utilizes a dedicated machine for each camera frustum. Previously improbable functions like rendering resolution, increasing frame rates, allows greater complexity in scene geometry and lighting than was previously possible.
This solution provides the ability to address the challenges of wider-angle lenses that require high resolution and multi-camera capture, which is taxing on current systems. This also means that the simplification of the signal chain reduces latency within the system.
