Whether FidelityFx Super Resolution (FSR) or the new Radeon Super Resolution (RSR), in both cases these are upscaling methods with the same goal: You should increase the FPS rates by pixel savings without the picture becoming unsightly. Thus, a lower resolution is calculated, the whole thing is highly scaled and the algorithm iron out as good as possible thereby emerging out.
As we know from past analyzes, it works more or less well according to the game and content. This works best with the NVIDIA Deep Learning Super Sampling process, which makes it possible to carry out temporal calculations and therefore has a greater impact. For this DLSS only works on RTX graphics cards. Both FSR and DLSS must be implemented by game developers. But what if that is not the case – so at thousands of games? Here only easier solutions help, that say: the driver must ensure that the pixel saving is smooth and unnoticed for them. That’s exactly what AMD has now introduced with Radeon Super Resolution.
Do not let the name lead to the mistake: Neither FSR Nor RSR operate really “Super Resolution” or “Super Sampling”, both methods save pixels to increase performance. A low resolution is always extrapolated on the native pixel amount of the connected monitor. What another upscaling procedure needs for when you can simply down the resolution in games? It simply looks a bit better. Modern liquid crystal screens only display an optimal, sharp image for native resolution. If a different, usually lower dissolution is set, visibly quality is lost. This problem can be dealt by the monitor with matching bobbins: reduce game or driver the resolution internally and let the monitor in faith that the native resolution arrives, better quality is achieved. Below than the native resolution is such upscaling despite compensating measures, such as reconstruction at a spatial level, aim is a relief of the graphics card and thus performance enhancement.
To home page
