During GDC 2022, Arkane Studios held a presentation titled A Guided Tour of Blackreef: Rendering Technologies in Deathloop to showcase the engine advancements featured in the game; the last section provided some interesting information about the upcoming FSR 2.0 implementation.
Deathloop will be the first game to support AMD’s new version of FidelityFX Super Resolution, which switches from a spatial scaling solution to a time scaling solution to dramatically improve game quality. ‘picture.
Indeed, Lou Kramer, Technology Developer Engineer at AMD, began his talk by pointing out with a screenshot comparison that the edges of the wooden pattern are very well reconstructed in the FSR 2.0 image (quality mode, from 1440p at 4K), to the point that they look better than the native 4K picture with TAA and Sharpener enabled.
Kramer then discussed FSR 2.0’s input sources in Deathloop, such as color buffer, depth buffer, and motion vectors. The input color buffer is in linear color space and the image format is R11G11B10_FLOAT. To improve precision in the target image format, the values are multiplied by a pre-exposure value. Since FSR 2.0 must also do this, the pre-exposure value is passed to it as a per-frame parameter. Deathloop also provides the exposure texture for the input, although this is not necessary as FSR 2.0 can calculate one.
Regarding output, Kramer said that Deathloop’s presentation is very smooth even when rendered natively, so there is a sharpener available in the graphics menu. However, FSR 2.0 has its built-in sharpener, RCAS (Robust Contrast Adaptive Sharpener); once FSR 2.0 is enabled, Deathloop’s is disabled to avoid excessive sharpness.
Deathloop is configured to support all modes of FSR 2.0 (from quality to ultra performance) and even dynamic resolution scaling.
The most interesting part of the presentation came last when we got a first look at the performance improvements that FSR 2.0 is expected to bring to Deathloop.
It should be noted that the numbers were saved on the beta versions, so they could be different in the final version. That said, compared to native 4K with TAA, Sharpening and Raytracing enabled, FSR 2.0 can increase performance by up to 50% in Quality mode (rebuilt from 1440p), up to 69% in Balanced mode (1270p), up to 90% in Performance mode (1080p) and up to 147% in Ultra Performance mode (720p). It’s on an AMD Radeon RX 6900 XT graphics card.
In another GDC 2022 conference entirely dedicated to FidelityFX Super Resolution 2.0, AMD engineer Thomas Arcila revealed the oversampling part of FSR 2.0 to be managed via Lanczos, the famous high quality interpolation filter. Ringing artifacts are avoided by clamping.
While the new version of FidelityFX Super Resolution will be cross-platform, AMD has spent time optimizing processes for RDNA 2 architectures. sampling them at runtime. According to AMD engineer Colin Riley, it’s faster on some RDNA 2 architectures.
Additionally, AMD has found that running certain FSR 2.0 shader workloads in Wave64 instead of Wave32 can improve runtime by up to 12% on a Radeon RX 6800XT. However, some of these RDNA 2 optimizations would run slower on non-RDNA 2 GPUs, including previous Radeon graphics cards and competing products.
In these cases the optimizations are marked and have a fallback implementation present so that they are disabled for certain GPUs. For example, the aforementioned WaveSize optimization is only active where performance would get an improvement.
Finally, AMD provided time estimates for FSR 2.0 integration. Games that already support NVIDIA DLSS 2.0 are the fastest (less than three days), while games without support for decoupled display/rendering resolution or motion vectors can take up to four weeks or more.
The first AMD FSR 2.0 games are expected to be available in Q2 2022. Stay tuned for more on FidelityFX Super Resolution 2.0 ahead of launch.