Checking if Nvidia's DLDSR offers native resolution with free performance.
Checking if Nvidia's DLDSR offers native resolution with free performance.
Thanks for the update on ReShade. I hadn’t heard of that filter before. I was already exploring CRU to tweak my monitor drivers, but it felt a bit confusing. I’m relieved you found what I needed and happy we got the right advice. It seems Nvidia might not want to support this approach because it could affect their business interests.
DLDSR captures a higher quality image and then reduces it for delivery. NIS, DLSS, and FSR work with lower quality images and boost them. It seems like a mix-up—didn’t I receive the right instructions? DLDSR/DSR starts from a native high-res source and gives you a lower-res version. That sounds more like downscaling rather than upscaling. If forced, I’d say upscaling would mean enlarging the input to match the output, which aligns better with what you’re describing. Consider starting with a large file and shrinking it for the final use. Let me know if that clears things up!
You're correct, DSR and DLDSR are techniques for downscaling purposes. More accurately, it's about increasing resolution significantly. DSDRS employ AI to upscale images, similar to how DLSS works (DL for Deep Learning). In fact, DLSS and DLDSR essentially refer to the same concept.
I understand you sometimes disagree with forum discussions, but this approach might have been taken too far. My perspective is solid, yet differences in how we frame things could matter. Would upscaling the effort rather than reducing it be more aligned? I’d love to hear your thoughts from the other two participants.
DSR and DLDSR are suitable for reducing resolution, while DLSS is effective for increasing it. However, it's possible to interpret this differently depending on the context.
The method remains consistent, but the display grid changes. Both approaches are designed for use at the screen's native resolution so they function correctly. The real advantage of DLDRS is that even with multipliers like 1.78 or 2.25, the image automatically scales to your native resolution thanks to the AI. If you apply the same multiplier without any prior scaling, the picture won't match the grid and will appear distorted despite the higher resolution.
DLDSR merely adjusts the image to match your screen's resolution without reducing quality. When I took a screenshot, the file was 3840x2160, shown on a 1440p monitor. It seems the image is rendered at full resolution initially and then downscaled before reaching the display. DLDSR always increases resolution beyond what was provided. DLSS actually performs downscaling, but artificial intelligence techniques make lower-res outputs look nearly identical to native quality.
It reverses the process of DLSS, which actually enlarges images to 1440p from lower resolutions. DLDSR does the opposite—it downsizes images to 1440p from higher resolutions. Your understanding is accurate but flipped. In DLDSR version 1.78, images remain at 1440p instead of being rendered higher. NVIDIA employs similar AI techniques in both DLSS and DLDSR to enhance quality, making them sophisticated AI-driven solutions. The 2.25 CG model upscales to 1920p and then converts it to 4K.