
























NVIDIA has published its latest update for the RTX Remix framework, with version 1.5 bringing impressive improvements. As some may recall, NVIDIA's RTX Remix is a tool for remastering older games to enable the game engine to support modern visual enhancements like full ray tracing, neural rendering with DLSS 4, and multi-frame generation. It also automatically enhances materials with generative AI. NVIDIA has been updating classic games like Half-Life 2 and Portal. According to the company blog post, RTX Remix 1.5 now includes new visual enhancements and a more efficient storage system that can reduce file sizes by roughly one-third, depending on the specific game.
The first major feature update is Smooth Normals. This feature, highly requested by the RTX Remix modding community, adds automatic smooth-normal generation to the legacy geometry in the game. This prevents older in-game assets from appearing blocky or faceted when path-tracing lighting is applied. Interestingly, NVIDIA has prepared a demonstration in a concrete tunnel within the game, which now looks like a uniform piece instead of the multi-piece rendering it once was. The company also added lighting control from various viewpoints, meaning that remixes and modding community will get to play with more lighting controls.
Last but not least, NVIDIA has updated its RTX IO storage technology with new algorithms and improved compression. Initially, RTX IO was developed as a high-performance storage technology to offload tasks from the CPU to the GPU, improving performance, speeding up game loads, and reducing CPU overhead. In-game textures now stream much faster, allowing gamers to notice a significant difference. One of the RTX IO updates is the file compression algorithm, which has been updated to support Portal with RTX, Portal: Prelude RTX, and the Half-Life 2 RTX games. In Portal with RTX, the installation size has decreased from 25 GB to 17 GB, while Half-Life 2 RTX has been cut from 80 GB to 50 GB.
此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。