The discussion reflects on the fifth anniversary of AMD’s FSR 1, acknowledging its initial shortcomings as a basic spatial upscaler that fell short of competing with Nvidia’s DLSS, but recognizing its role as an early step towards more advanced upscaling technologies like FSR 2 and beyond. Despite its limitations and the rise of superior machine learning-based solutions, FSR 1 remains in use in some games due to its ease of integration, illustrating the ongoing evolution and varied adoption of upscaling methods in the gaming industry.
The discussion marks the fifth anniversary of AMD’s FSR 1 (FidelityFX Super Resolution 1), reflecting on its impact and evolution over the years. Initially, FSR 1 was met with criticism, particularly from figures like Alex, who openly expressed disappointment with its quality. At the time, it was a basic spatial upscaler with sharpening, not designed to compete with Nvidia’s DLSS 2, which had already set a high standard for temporal upscaling. Despite its shortcomings, FSR 1 was AMD’s early attempt to offer an alternative upscaling technology, but it fell short in image quality and performance expectations.
The conversation highlights that while FSR 1 was a significant step for AMD, it was never intended to be a direct competitor to DLSS. Early adopters and reviewers found it lacking, and it even required testing on games like Godfall, which has since faded from prominence. The real advancements came with FSR 2, which introduced temporal upscaling and marked a substantial improvement, though it still wasn’t quite a “DLSS killer” as some had hoped. Subsequent iterations, including FSR 3 and FSR 4, have continued to close the gap, enhancing image quality and performance, though the technology still doesn’t dominate the market.
Interestingly, FSR 1 remains in use even five years later, especially in certain Nintendo and Capcom titles, such as those running on the RE Engine. Developers appear to favor FSR 1 for its ease of integration and spatial upscaling, despite the availability of better temporal upscaling solutions that generally provide superior visual results. The panelists note that FSR 1 tends to produce a “painterly” or softened image that some find unattractive, sometimes preferring simpler bilinear upscaling over it. This continued use is somewhat puzzling given the availability of more advanced options, but it may be tied to cost and development convenience.
The discussion also touches on the broader context of upscaling technologies in gaming, noting that machine learning and temporal upscaling are becoming the dominant trends. While FSR 1 was open source and accessible, newer approaches, including Intel’s XeSS and AMD’s later FSR versions, leverage machine learning to deliver better results. The conversation reflects on how the industry is converging on these more sophisticated methods, even as some legacy technologies like FSR 1 persist in certain titles, sometimes to the detriment of visual quality.
Finally, the hosts share some nostalgic remarks about revisiting past gaming technology milestones through resources like the Wheel Jad website, which archives gaming news and performance data from previous E3 events and other milestones. They express enjoyment in looking back at these stories and anecdotes, highlighting the rapid evolution of gaming tech over the past decade. The retrospective serves as a reminder of how far upscaling and graphical enhancement technologies have come since the early days of FSR 1, setting the stage for future innovations in the field.