AMD FidelityFX Super Resolution goes open source, more games supported.

AMD FidelityFX Super Resolution goes open source, more games supported.

0

Images in this article provided by: AMD (https://www.youtube.com/watch?v=Br5dDyHMoFk)

Less than a month after FidelityFX Super Resolution (FSR) was first introduced in a select handful of titles, it’s already gone fully open source!

Modding potential

This is particularly interesting for mod support. Even before today’s announcement, we’ve already seen FSR modded into high-profile games like Grand Theft Auto 5. With the technique opened up for anyone, it allows for further experimentation and unofficial, user-created support.

Moving into the future, FSR’s likely to be included in the majority of upcoming releases. But what about pre-existing games not receiving long-term updates? Hopefully, the tech-savvy PC gaming community will step up and create their own FSR intergrations.

Four more FSR games

Even more games are receiving FSR as time goes on. Resident Evil 8 will be the first ‘true’ AAA game to support FSR. We’re looking forward to seeing how capable the technique is in a big-budget blockbuster.

Intriguingly, two of the announced games (Edge of Eternity and Necrominda: Hired Gun) already have bespoke DLSS implementations. In our understanding, this will be the first time we’ll be able to compare the technique’s outputs side-by-side in a like-for-like scenario.

We’re fully expecting DLSS to emerge victorious, but the question is by how much? Let’s quickly cover the techniques.

DLSS VS. FSR

DLSS and FSR are two very different approaches to producing high resolution frames from lower resolution inputs and should be judged accordingly.

In basic terms FSR simply upscales the frame, whereas DLSS reconstructs it.

DLSS is integrated right into the game engine. It’s temporal based, and accumulates information over previous frames, to re-construct frames to a higher perceived resolution.

It sounds like the optimal solution, but unlike FSR, requires a more hands-on approach from developers to get it up and running. FSR is practically drag-and-drop in comparison.

DLSS is also a closed-source, proprietary technology. A ‘black box’. You feed it information and it magically spits out a result. No peeking behind the curtain to see how it works. This limits potential in tweaking, implementing and porting the technique.

Since DLSS meddles with previous frames, it produces some unintended side effects. Ghostly trails and blur were most notable in earlier iterations of DLSS, but NVIDIA has been consistently updating DLSS to minimise artifacts.

Red Dead Redemption 2, Rockstar Games’ crowning achievement, was recently updated with DLSS. Though, its results show all implementations aren’t created equal.

The benefits of FSR

FSR is primitive in comparison. A ‘spatial’ post-processing shader which looks at an already anti-aliased frame and upscales with no data to leverage.

FSR’s limited capabilities are clear, but its simplicity is also its greatest strength.

To compared both the technique’s launch windows, FSR is clearly easier and faster for developers to implement. Unlike DLSS, it doesn’t rely on AI-powered Tensor Cores or key data like history buffers or motion vectors, complicating the process.

The full shader source code is provided by AMD, with wide support for the most popular APIs like DirectX12, DirectX11 and Vulkan.

For gamers, FSR’s paired back hardware requirements means it practically runs on anything. It even works on APUs with integrated graphics.

Even though it’s an AMD technology, it’s not arbitrarily locked to their ecosystem. AMD GPU, NVIDIA GPU or even a games console – FSR is platform-agnostic.

AMD has developed a bunch of useful technologies to boost performance, like Smart Access Memory.

If you’re struggling to lock to 60 frames-per-second on an aging graphics card, don’t upgrade just yet. FSR may be enough to boost performance for a perfectly paced experience.

The main concern regarding FSR is whether it’s even worthwhile. We’re in a post-resolution era, and game developers have been slowly iterating on their own temporal-based reconstruction techniques for a while now.

DLSS is unique because it leverages specialised hardware and NVIDIA’s AI expertise. But what about FSR, why shouldn’t developers just use their own proven, in-house techniques?

Engine integration

Most excitingly, the announcement also reveals upcoming compatibility with big-name engines like Unity and Unreal Engine. Game development is slowly converging on these two colossal titans of the industry. Boot up a game, and you’ll likely encounter their logos on the splash screen.

This will allow small-time indies all the way up to AAA developers using Unity and Unreal Engine to enable FSR in their games.

FSR may be key to squeezing out as much power from lower-end systems, like the recently announced Steam Deck, a handheld gaming PC.