What is real-time 3D?

The main characteristic of real-time 3D is its interactivity with immediate response. Users can interact with the 3D environment, move objects, adjust settings, or even create new elements in real-time. Depending on the actions they perform, results are generated nearly instantly (around 1/60th of a second per frame), providing highly natural interactivity.

Real-time 3D systems are designed to display 3D images at high refresh rates, often ranging from 30 to 60 frames per second (FPS). This enables a smooth, low-latency user experience.

Unlike pre-rendered 3D, where images are rendered in advance (for videos or image sequences), real-time 3D generates images on the fly, performing complex calculations in real-time for each displayed frame.

Real-time 3D is used in a wide range of applications, from video games, where it enables players to control characters and explore virtual worlds, to product design, where it facilitates the visualization of 3D prototypes, as well as in training, simulation, and many more.

At SoWhen?, we use real-time 3D to meet our clients’ interactivity needs through custom virtual worlds, phygital experiences with behavioral AI, learning games, and phygital models to convey scientific concepts to the general public.

Here are some examples of projects carried out using Unreal Engine (UE5): Alstom’s virtual universe, Michelin’s 3D film shown during the 24 Hours of Le Mans centenary, Disney+’s interactive experience ‘The Wall of Shadows,’ and the portable phygital model by Regen Lab.

To generate complex worlds with high fidelity lighting and materials, featuring extensive data and a large variety of interactions, it is often necessary to have powerful computer hardware, especially high-performance graphics cards (GPUs). These expensive machines now have a more flexible and accessible digital counterpart through cloud computing.

According to the Unreal Engine (UE) Forrester report, 77% of companies aim to enhance team collaboration using real-time tools. Epic Games has therefore decided to increase UE integration into a cloud-focused ecosystem (web-hosted virtual machines) with features like Pixel Streaming. This allows companies to address their specific needs by adjusting subscriptions instead of purchasing physical machine clusters.

In the beginning, real-time rendering couldn’t achieve the same image quality as pre-rendered 3D.

Over the past 40 years, computational power has overcome certain constraints, enabling the generation of highly realistic real-time images.

The quality gap between the two is narrowing with technological advancements, leading to an ever-expanding range of applications and industries for real-time rendering.