AMD’s Multi Frame Generation Feature May Be Coming Soon

An intriguing detail found in AMD’s software tools has sparked excitement, as the Multi Frame Generation technique, which helped NVIDIA gain an edge in smoothness, seems to be on the verge of arriving on Radeon.

The signal is concrete, with a new “ratio” option linked to image generation appearing in the FidelityFX ecosystem, aligning with the idea of multiplying artificial images between two rendered images. AMD currently remains limited to a 2x image generation while NVIDIA already offers more aggressive modes. While not a guarantee of a release date, it resembles the period before a feature transitions from the lab to developers’ kits. The question now is whether AMD’s MFG will finally close the gap or simply reduce it.

AMD Introduces an “Upgrade Ratio” Option in FidelityFX SDK

A major clue is the addition of a new feature in the SDK, labeled with a lengthy but clear name, an “ratio” option for image generation. The stated goal is to allow users to choose an “optimal” multiplier between performance and visual quality. In gaming terms, this means you are no longer limited to a single frame generation mode, allowing you to aim for a higher level based on your GPU margin and game rendering.

This development is crucial as this type of setting is only meaningful if AMD offers multiple tiers, something that competitors are already doing. Currently, AMD mainly offers a limited schema, with only one frame generated between two native frames, equivalent to FSR 4 at 2x. The addition of a “ratio option” hints at moving towards higher modes, potentially 3 or more generated frames, although the exact numbers are not yet public.

In the developers’ daily routine, such a feature rarely arrives just for show. When an API exposes a ratio parameter, it means testing behaviors, validating artifacts, measuring latency, and adjusting profiles based on different scenes need to be done. An engine engineer, “Marc,” simplifies it by stating that if you are given a ratio selector, it is expected that you will use it, or else it wouldn’t have made the cut. In other words, AMD is laying the groundwork for a broader deployment of Multi Frame Generation.

NVIDIA DLSS 4.5 and Intel XeSS 3 Have Already Set the Bar

The tech specs showcase AMD’s lag behind, as NVIDIA has been using image generation since DLSS 3 in 2022 and then advanced further with multi-image generation on the RTX 50 series. The concept involves rendering a native image and inserting multiple synthetic images between two renders. In recent implementations, NVIDIA offers modes up to 6x, with dynamic adjustment possible based on the screen refresh rate.

On the other hand, Intel has also embraced MFG through XeSS 3, offering modes up to 4x on Arc, including on some integrated chips. This shift in the market perception happens when two out of three manufacturers already offer multi-frame technology, leaving the third perceived as “lagging behind,” even as pure upscaling has improved. In demanding games, especially those utilizing options like path tracing, the race for smoothness becomes a clear marketing point.

However, a word of caution is required. While multi-frame technology can enhance FPS display, it is not without cost. The more generated frames, the higher the risk of faults on fast movements, particles, HUD elements, or small inconsistencies on highly contrasted scenes. “Marc” bluntly states that while 6x may impress, it can also reveal strange artifacts during camera rotations. Therefore, AMD will be judged on a specific point – visual stability, not just numbers.

How Multi Frame Generation Could Change the Game for Radeon Cards

If AMD fully integrates MFG into its suite, the most immediate impact will be on the perception of smoothness, especially for players aiming to match a 120Hz or 144Hz refresh rate closely. The idea of a selectable ratio opens up the possibility of different profiles, offering a balanced mode to avoid artifacts and a more aggressive mode for competitive games where a consistent animation is crucial, even if the image is not perfect under scrutiny.

Another consequence is the developers’ implementation. With the ratio parameter exposed in tools, studios can provide a finer menu or at least a simple setting like 2x, 3x, 4x, based on what the pipeline supports. This could also facilitate “dynamic” approaches, where the game adjusts the ratio based on the workload, such as transitioning from a narrow corridor to a vast open area. The aim is to prevent FPS fluctuations without compromising on latency.

Yet, one aspect to monitor is that AMD has not provided a precise public timeline for these technical elements. Clues suggest an “imminent” rollout in the 2026 timeframe, but it does not specify whether it will be initially activated on a few recent Radeon models or broader. The true test will be in games – rapid adoption by three or four major titles will have an immediate effect, otherwise, the feature may remain a line in a changelog that few people utilize.