NVIDIA’s Reflex technology just got a huge upgrade, . The company’s Reflex 2 improves things several times over, thanks to some software skills. The end result? It promises to further reduce gaming latency by up to 75 percent. In this case, latency refers to how quickly the game displays interaction from player input.
Here’s how it works, and it’s pretty wild. Reflex 2 combines the pre-existing low latency mode with something called Frame Warp. This technology tasks the CPU to calculate the position of the next frame as the current frame is rendered by the GPU. The CPU uses player input, usually via the mouse, to calculate the next camera position.
The tools then sample the next camera position and graft it onto the current frame, waiting until “the last moment to ensure the most recent input appears on screen.” This glimpse into the near future should “allow players to improve their aim and track enemies.”
But how can a previously rendered image exist on top of the current frame without creating all sorts of visual errors? NVIDIA has developed a “latency-optimized predictive rendering algorithm” that uses data from previous frames to fill in any obvious gaps, so players see a new frame with updated camera position without any weird gaps. It is, literally, a distortion of the frame (hence the name).
The company demonstrated this technology with Embark Studio’s Finalswhich will be one of the first games to support the toolkit. The proof is in the pudding. With the game running on an RTX 5070 GPU at 4K resolution with maximum settings, latency was 56ms. The first iteration of Reflex reduced the latency to 27 ms, while Reflex 2 reduced it to 14 ms.
In addition FinalsNVIDIA says so brave Will soon support technology. The company notes that it used Reflex 2 with Frame Warp to reduce latency to an average of less than 3 ms. Reflex 2 will be shown for the first time as part of with support for other RTX GPUs coming at some point in the future.
If you buy something through the link in this article, we may earn a commission.
https://s.yimg.com/ny/api/res/1.2/jj5kF667bgMGVWWnFvAUVg–/YXBwaWQ9aGlnaGxhbmRlcjt3PTEyMDA7aD02NDk-/https://s.yimg.com/os/creatr-uploaded-images/2025-01/2347e810-cd1f-11ef-abfe-0c034fd25d55
Source link