Microsoft adds Auto Super Resolution to more games for Copilot+ PCs

Microsoft has introduced a significant enhancement for PC gaming with the integration of Auto Super Resolution (Auto SR) into Copilot+ PCs. This AI-powered technology aims to boost frame rates and improve visual fidelity in a growing number of games, offering a smoother and more detailed gaming experience for users of these new devices.

The Auto SR feature is designed to work seamlessly with Windows, leveraging the dedicated Neural Processing Unit (NPU) found in Copilot+ PCs, particularly those equipped with Snapdragon X processors. By dynamically lowering a game’s rendering resolution and then employing sophisticated AI algorithms to upscale the image, Auto SR enhances visual quality while simultaneously increasing performance, a dual benefit that can transform gameplay.

Understanding Auto Super Resolution

At its core, Auto Super Resolution is an AI-driven upscaling technology. It operates by rendering games at a lower resolution to gain a performance advantage, then utilizes a convolutional neural network (CNN) expertly trained on gaming content to reconstruct and enhance the image to a higher definition. This process is managed by the NPU, a specialized processor designed for AI tasks, which offloads the computational burden from the CPU and GPU. This offloading allows the CPU and GPU to focus more on rendering the game itself, leading to increased frame rates and a more fluid visual experience.

The technology is integrated directly into Windows, aiming for a low-friction user experience without requiring extensive manual configuration or developer intervention for supported titles. This OS-level integration means that Auto SR can be applied broadly, enhancing existing games in the player’s library. Unlike some other upscaling solutions that require specific game engine integration, Auto SR works by processing the completed game frame.

Microsoft emphasizes that Auto SR is designed to provide a balance between visual quality and performance. It aims to deliver visuals that surpass native 1080p quality while maintaining the fast frame rates typically associated with lower resolutions. The AI model is trained to add detail and perform anti-aliasing during the upscaling process, contributing to a sharper and more refined image.

Technical Implementation and Requirements

For Auto Super Resolution to function, specific hardware and software prerequisites must be met. The technology is currently exclusive to Copilot+ PCs, which are defined by their inclusion of a Qualcomm Snapdragon X or X2 Series processor featuring an integrated Hexagon NPU. These NPUs are crucial for the AI computations required by Auto SR, providing the necessary processing power with high efficiency.

Beyond the processor, a Copilot+ PC must be running Windows 11, version 24H2 or later, and have the latest graphics and neural processor drivers installed. A display resolution of 1080p or greater is also a requirement for Auto SR to operate effectively. The system also needs the latest version of the Auto Super Resolution Package, which is typically managed through Windows Update.

Game compatibility is another key factor. Auto SR is supported in most games running on DirectX 11 or DirectX 12, and can work with emulated x64 or Arm64 native platforms. However, it’s important to note that Auto SR may not be compatible with games using certain formats, such as 10-bit formats, or those running on other graphics runtimes like DirectX 9, Vulkan, or OpenGL. The input resolution for Auto SR is also generally limited to a range between 700 and 900 pixels tall, meaning that games may need to be configured to an Auto SR-compatible resolution, such as 1280×800, for the feature to work optimally.

How Auto Super Resolution Enhances Gaming

The primary benefit of Auto SR is its ability to significantly boost frame rates in games, making them run more smoothly and responsively. This is achieved by lowering the game’s rendering resolution, which reduces the workload on the graphics processing unit (GPU). The AI then intelligently upscales this lower-resolution image, restoring and even enhancing visual detail that might otherwise be lost.

This dual action—reducing rendering load and intelligently upscaling—allows for a higher frame rate without a drastic compromise in visual quality. Microsoft has demonstrated instances where Auto SR can improve frame rates by up to 60 percent. For example, in Borderlands 3, Auto SR was shown to increase frame rates from 26 fps to a more playable 41 fps while maintaining acceptable image quality.

The technology aims to provide an experience that is visually comparable to native resolutions but with the performance benefits of lower ones. Users can expect a smoother, more detailed gaming experience, particularly in titles that are graphically demanding or on hardware that might otherwise struggle to maintain consistent frame rates. The AI’s ability to add detail and perform anti-aliasing during the upscaling process contributes to a cleaner and more visually appealing final image.

Automatic vs. Manual Configuration

Auto SR is designed with a focus on automatic enhancement, aiming to provide a seamless experience for players. For a curated list of games verified by Microsoft, Auto SR is enabled by default. Upon launching these supported titles, users will typically receive a notification confirming that Auto SR is active, requiring no further input.

However, the feature also offers flexibility for users who wish to apply it to games not automatically supported. This can be done on a per-game basis through Windows Graphics Settings, where users can opt-in specific applications. When opting in games manually, their custom Auto SR settings take precedence over the default. To enable Auto SR for such games, users need to navigate to Settings > System > Display > Graphics, locate the game’s executable file, and then select the option to turn on Auto SR.

It’s important to note that manually opted-in games do not follow the global Auto SR default settings. Users must manage these settings individually for each game they wish to enhance. Furthermore, if a game is set to exclusive fullscreen mode, the in-game resolution may need to be adjusted to an Auto SR-compatible setting. For borderless windowed fullscreen, changing the desktop display resolution to an Auto SR-compatible setting before launching the game might be necessary.

Supported Games and Future Expansion

At launch, Microsoft identified a specific set of games that would work with Auto SR out-of-the-box. This initial list included titles such as BeamNG.drive, Borderlands 3, Control (dx11), Dark Souls III, God of War, Kingdom Come: Deliverance, Resident Evil 2, Resident Evil 3, Sekiro: Shadows Die Twice, Sniper Ghost Warrior Contracts 2, and The Witcher 3. A third-party, open-source website, “Works on Windows on Arm,” collaborates with Microsoft to maintain an evolving list of games compatible with Auto SR, allowing users to filter for games with “Yes, out-of-the-box” or “Yes, opt-in” support.

While the initial rollout focuses on a curated list, Microsoft’s intention is for Auto SR to become more widely applicable. The technology is designed to enhance existing games without requiring developer modifications, suggesting a broad potential for future compatibility. As more Copilot+ PCs become available and the technology matures, the number of games that Auto SR can enhance is expected to grow significantly.

The development of Auto SR is part of a broader trend in PC gaming where AI upscaling technologies are becoming increasingly prevalent. While Auto SR is currently exclusive to Snapdragon X-powered Copilot+ PCs, future iterations or similar technologies from Intel and AMD may bring comparable AI-driven enhancements to a wider range of hardware.

Potential Limitations and Considerations

While Auto Super Resolution offers compelling benefits, there are a few considerations and potential limitations to keep in mind. One aspect is the introduction of a small amount of latency. Microsoft claims that Auto SR adds, on average, only a single frame of latency, which is often imperceptible to most players. However, in competitive gaming scenarios where every millisecond counts, this added latency might be a factor for some users.

Auto SR primarily functions as a screen-space upscaler, meaning it works with the base image provided by the game. Unlike some other upscaling technologies, it doesn’t have access to motion vectors or depth information from the game engine itself. This can sometimes lead to artifacts, particularly with text elements or HUDs, which might appear less crisp than other parts of the image. Additionally, Auto SR does not currently work with HDR displays, which could be a drawback for users with OLED or high-end IPS panels who prioritize color vibrancy and accuracy.

The technology also has specific requirements regarding input resolution, typically needing games to render below 900 lines vertically. This might necessitate manual adjustment of in-game resolution settings for games that do not automatically configure to a compatible mode. Furthermore, Auto SR may not support all DirectX 11/12 games, especially those employing certain formats like 10-bit color depth, or games running on older graphics APIs.

The Role of NPUs in AI Upscaling

The integration of the Neural Processing Unit (NPU) is central to Auto Super Resolution’s functionality and efficiency. NPUs are specialized hardware components designed to accelerate artificial intelligence workloads, such as the complex calculations required for machine learning models. By dedicating the AI upscaling tasks to the NPU, Auto SR offloads this processing from the CPU and GPU.

This offloading strategy is particularly beneficial for Copilot+ PCs, which often feature integrated GPUs that may not have the raw power of high-end discrete graphics cards. The NPU handles the heavy lifting of the AI model, allowing the GPU to focus on rendering frames at a higher rate. This synergy ensures optimal visual fidelity and peak performance with minimal overhead, as the NPU is optimized for these specific types of operations.

The use of an NPU also contributes to better power efficiency. AI tasks can be run on a low-power section of the chip, which is advantageous for battery life in laptops. This dedicated hardware approach contrasts with some other upscaling technologies that rely more heavily on the GPU’s general-purpose processing cores, potentially consuming more power and impacting overall system performance.

Comparison with Other Upscaling Technologies

Auto Super Resolution enters a landscape already populated by established AI upscaling technologies like NVIDIA’s DLSS (Deep Learning Super Sampling), AMD’s FSR (FidelityFX Super Resolution), and Intel’s XeSS (Xe Super Sampling). While all these technologies aim to improve performance and visual quality by upscaling lower-resolution images, Auto SR distinguishes itself through its OS-level integration and automatic application to existing games without developer input.

Unlike DLSS, FSR, and XeSS, which typically require integration into the game’s engine and often leverage motion vectors and temporal data for enhanced stability, Auto SR operates as a post-process on the rendered frame. This means it has less data to work with, potentially leading to differences in image quality and artifacting compared to solutions that have deeper access to game engine data. Auto SR’s reliance on the NPU also differentiates it from technologies that primarily utilize the GPU.

Microsoft’s approach with Auto SR is to democratize AI upscaling for a wider range of games and users, particularly on their new Copilot+ PCs. While DLSS and FSR often provide more granular control and potentially superior image quality due to their deep integration, Auto SR offers a more accessible, plug-and-play experience for a broad library of titles. However, the quality of Auto SR’s output is still being evaluated against these more mature technologies, with some comparisons suggesting it performs better than basic bilinear upscaling but may not match the visual fidelity of DLSS.

The Future of AI in PC Gaming and Copilot+ PCs

The introduction of Auto Super Resolution on Copilot+ PCs signals a significant step towards AI becoming an integral part of the PC gaming experience. As AI capabilities continue to advance, technologies like Auto SR are poised to become more sophisticated, offering even greater improvements in performance and visual fidelity.

Microsoft’s investment in AI-driven features for its PCs, including Auto SR, highlights a strategic shift towards leveraging specialized hardware like NPUs. This trend is expected to continue, with future processors from Intel and AMD also incorporating advanced NPUs, potentially expanding AI upscaling capabilities to a broader ecosystem of devices. The focus on AI also aligns with the growing demand for more immersive and visually stunning gaming experiences, even on portable or less powerful hardware.

As AI upscaling matures, it may address current limitations such as latency and artifacting, becoming an even more indispensable tool for gamers. The ongoing development and refinement of these technologies suggest a future where AI plays a pivotal role in ensuring that games run smoothly and look their best across a wide range of hardware configurations.