Neural Super-Resolution for Real-Time Rendering with Radiance Demodulation

Figure 1. Quality comparison between our method and the state-of-the-art methods NSRR [53], BasicVSR++ [5], TTVSR [32] and RVRT [29]. The upsampling ratio is set as 4 × 4. By radiance demodulation, we perform super-resolution on the smooth lighting component only, allowing our method to preserve richer scene details after re-modulation with the high-resolution material component.