Latest news Software

AMD extends AFMF 2.1 with “Fast Motion Response” – finally an answer to fast game scenes?

AMD has added a new feature to its frame generation technology AFMF (Fluid Motion Frames) in version 2.1, which goes by the name of “Fast Motion Response”, a technical retrofit kit that is basically intended to solve an old problem, report our colleagues from WccfTech: motion blur and artifacts in fast game scenes. However, the whole thing is not yet available in the normal Adrenalin software, but only in the preview driver “PyTorch on Windows” (version 25.120.01.14) and only works with GPUs from the Radeon RX 7000 and RX 9000 series.

While the official Adrenalin edition 25.9.2 only offers two modes under AFMF 2.1 “Search Mode” and “Performance Mode”, the PyTorch version brings two new options into play: “Repeat Frame” and “Blended Frame”. What sounds unspectacular at first glance is technically a late but necessary addition to mitigate the weaknesses of previous frame generation for fast movements. The way it works can be summarized as follows: “Repeat Frame” mode simply repeats the last valid frame instead of generating a potentially erroneous new frame. This is particularly useful for quick changes of direction, camera pans or action scenes, in which classic frame interpolation often leads to ghosting, artifacts or even “distorted images”. The image quality remains constant, even if the movement does not appear quite as smooth.

The second option, “blended frame”, takes a different approach: two adjacent frames are blended to approximate an image in between. This can subjectively lead to a smoother display, but creates visible blurring depending on the speed of movement, a compromise that is either welcome or annoying depending on the style of play. It is interesting that with this function, AMD is apparently finally overcoming the restriction that previously served as a stopgap solution in “Search Mode”: This completely deactivated frame generation if movements were classified as “too fast”. With “Fast Motion Response”, AFMF can now also remain active in such scenes, albeit with an individually selectable quality/performance compromise.

Overall, the extension seems like a long overdue patch for a feature that previously shone in slow to moderate game scenes, but was often useless or even annoying in fast titles. The fact that AMD has not (yet) officially integrated the function into the main drivers is probably due to its experimental nature and the desire for feedback from the community. Nevertheless, the direction is clear: with AFMF 2.1, AMD wants to be on a par with NVIDIA’s DLSS 3.5 without falling into the traps of the early frame generators. However, the dependence on the PyTorch preview and the limitation to current GPU generations also show that AMD no longer wants to drag along old hardware. If you want to benefit from the new functions, you need modern cards and apparently also patience during the rollout.

We can only hope that AMD will soon release a final implementation that will also reach mainstream users. Because the potential of AFMF – especially with regard to the new options – is high. But without stable, widely available integration, it will remain an insider feature for hobbyists and early adopters.

Source: WccfTech

Bisher keine Kommentare

Kommentar

Lade neue Kommentare

Redaktion

Artikel-Butler

3,727 Kommentare 14,339 Likes

AMD hat seiner Frame-Generation-Technologie AFMF (Fluid Motion Frames) in Version 2.1 ein neues Feature spendiert, das auf den Namen „Fast Motion Response“ hört, ein technischer Nachrüstsatz, der im Grunde ein altes Problem lösen soll, berichten unsere Kollegen von WccfTech: Bewegungsunschärfe und Artefakte bei schnellen Spielszenen. Das Ganze ist bislang allerdings nicht in der normalen Adrenalin-Software verfügbar, sondern nur im Preview-Treiber „PyTorch on Windows“ (Version 25.120.01.14) und funktioniert ausschließlich mit GPUs der Serien Radeon RX 7000 und RX 9000. Während die offizielle Adrenalin-Edition 25.9.2 lediglich zwei Modi unter AFMF 2.1 bietet „Search Mode“ und „Performance Mode“ bringt die PyTorch-Variante zwei neue […] (read full article...)

Antwort Gefällt mir

Danke für die Spende



Du fandest, der Beitrag war interessant und möchtest uns unterstützen? Klasse!

Hier erfährst Du, wie: Hier spenden.

Hier kannst Du per PayPal spenden.

About the author

Samir Bashir

As a trained electrician, he's also the man behind the electrifying news. Learning by doing and curiosity personified.

Werbung

Werbung