Source: NVIDIA
Advanced 3D Rendering Concepts
NVIDIA continues to push the boundaries of rendering technology, moving towards a future where fewer frames are traditionally rendered. With the introduction of DLSS 4.5, the frame generation capability has increased significantly. Building on the 4x Frame Generation from the previous generation, DLSS 4.5 now offers up to 6x frame generation. This means that for every single frame natively rendered, up to five additional frames can be generated. Further details are available on the official NVIDIA page.
DLSS 4.5 brings a substantial improvement in visual fluidity through its new 6X Mode of Multi Frame Generation. Utilizing a second-generation transformer model for enhanced image stability, along with advancements in frame pacing and overall image quality, DLSS 4.5 can produce five extra frames for each natively rendered frame, achieving a maximum 6X multiplier. This innovation is designed to provide the smoothest experience for path-traced games, fully leveraging high-refresh-rate monitors like 240Hz and 360Hz displays. On GeForce RTX 50 Series GPUs, upgrading from 4X to 6X Multi Frame Generation can boost 4K frame rates in path-traced titles by as much as 35%. When combined with NVIDIA Reflex low latency technology, DLSS 4.5 generates these additional frames with minimal impact on responsiveness.
Additionally, DLSS 4.5 introduces Dynamic Multi Frame Generation. NVIDIA likens this intelligent system to an automatic transmission in a vehicle, as it automatically adjusts performance to suit various system demands.
Dynamic Multi Frame Generation in DLSS 4.5 is an intelligent system that operates like an automatic transmission for a GPU. Instead of maintaining a fixed multiplier, it dynamically shifts between different frame multipliers to achieve an optimal balance of frame rate, image quality, and responsiveness. This means it only generates the necessary frames to maximize the target frame rate or the display’s refresh rate, whether it’s 120Hz, 144Hz, 240Hz, or higher.
The system continuously monitors the difference between the GPU’s performance and the display’s maximum refresh rate, adapting in real-time. In graphically demanding situations, it ‘upshifts’ to increase frame generation, compensating for performance drops and ensuring smooth visuals on high-refresh displays. Conversely, when the workload decreases, it seamlessly reduces the multiplier, only computing what is essential.
