Nvidia released the DLSS 3.1.13 version update yesterday. In addition to fixing the bugs in the previous version and improving stability, the biggest highlight is the improvement of deep learning anti-aliasing (DLAA) for developers.
After the game developers upgrade the new version of DLSS, they can add DLAA mode in the DLSS menu, so as to reach 5 modes.
Five DLSS operation modes are as follows:
Ultra-Performance Mode: Maximum performance output
Performance Mode: high performance output
Balanced Mode: Balance game quality and performance
Quality Mode: focus on quality
DLAAIn the SDK update description, the unused G preset is also introduced, and the official description is not provided.
Deep Learning Anti-Aliasing is a form of spatial anti-aliasing created by Nvidia. DLAA relies on and requires Tensor Cores available in Nvidia RTX cards.
The anti-aliasing method of DLAA is similar to deep learning super sampling, and one important difference is that the goal of DLSS is to improve performance at the expense of image quality, while the primary task of DLAA is to improve image quality at the expense of performance, which is different from resolution upscaling or Minification is irrelevant.
