NVIDIA DLSS 3.5 announced
A new upscaling technology has emerged, which is essentially an enhancement to NVIDIA’s existing DLSS3.
While there haven’t been any updates regarding AMD FSR3, and XeSS going open-source, NVIDIA is determined to make headlines with the latest iteration of DLSS (Deep Learning Super Sampling). This AI-driven algorithm has undergone four significant releases to date: DLSS1 (March 2020), DLSS 2.3 (November 2021), DLSS 3 (October 2022), and DLSS 3.1 (January 2023). The most recent update appears to be focused on improving ray tracing.
DLSS 3.5 is focusing in on Ray (Tracing) Reconstruction (RR), delivering a visibly enhanced visual experience compared to traditional denoising techniques. NVIDIA claims that Ray Reconstruction has been trained on five times more data than DLSS3. It has been trained to incorporate additional game and software engine data, recognize various ray-traced effects, distinguish between good and bad temporal and spatial pixels, and preserve high-frequency data for upscaling.
The utilization of denoisers was found to strip away essential data required for upscaling, leading to a loss of color data during denoising and subsequent upscaling. Furthermore, traditional methods may result in inaccurate lighting effects due to the accumulation of pixels from previous frames, often resulting in ghosting.
Denoisers also contribute to subpar global illumination and lower-quality reflections, as there isn’t enough data for interpolation. This will not be the case with Ray Reconstruction, which is the primary feature of DLSS 3.5 update.
NVIDIA has confirmed that DLSS 3.5 will make its debut in the fall, featuring in titles such as Cyberpunk 2077: Phantom Liberty, Portal with RTX, and Alan Wake 2. Additionally, it will be available in the NVIDIA Omniverse Platform, Chaos Vantage, and D5 Renderer. The “RR” tech will work across all RTX GPUs (unlike Frame Generation).