Control: Resonant Hands-On: DLSS 4.5 Transforms Ray-Traced Manhattan
Remedy Entertainment has built a distinctive identity around surreal storytelling, supernatural action, and meticulously designed environments. With Control and Alan Wake forming the foundation of the Remedy Connected Universe (RCU), the studio has created a shared world where narrative, Brutalist architecture, and supernatural phenomena reinforce one another.
Released in 2019, Control quickly became a technical and artistic showcase. It won IGN’s Game of the Year award, received eight nominations at The Game Awards 2019, and ultimately won Best Art Direction. The game also became one of the early showcases for real-time ray tracing and DLSS on PC.
Now, Control: Resonant takes that foundation into a much larger and stranger version of Manhattan. The protagonist changes, the combat system becomes more melee-focused, and the rendering pipeline moves to DLSS 4.5, Path Tracing, and Ray Reconstruction.
Ahead of its September 24 launch, our hands-on session offered a look at how Remedy is combining gameplay changes with modern AI-assisted rendering to create a new interpretation of the New Weird.
⚔️ A New Faden Takes Center Stage #
Unlike the original Control, which placed Director Jesse Faden at the center of the story, Control: Resonant follows her younger brother, Dylan Faden.
Players previously encountered Dylan as a mysterious figure held by the Federal Bureau of Control (FBC). In Control: Resonant, he finally becomes the protagonist, bringing a fundamentally different perspective to the supernatural world established by the original game.
The setting also changes dramatically.
Instead of being confined primarily to the Oldest House, the new game moves into a distorted Manhattan transformed by invasive cosmic forces. Gravity, physical laws, lighting, and architecture no longer behave according to familiar rules.
That shift in setting also changes how combat works.
Melee Becomes the Core Combat Mechanic #
Where the original Control emphasized supernatural abilities and firearm-based combat, Control: Resonant puts close-quarters combat much closer to the center of the experience.
Dylan’s primary supernatural weapon is a shape-shifting weapon called Aberrant, which can transform between different forms and support distinct combat styles. Combined with supernatural abilities and talents, the weapon system is designed to make player choices an important part of character progression.
The progression system is also built around replayability.
Skills, weapon upgrades, and talents change according to player decisions, and the complete system cannot be unlocked during a single playthrough. Players who want to explore the full range of Aberrant’s capabilities will therefore need multiple runs.
The main story is expected to take roughly 30 hours, while a more comprehensive completion run can reach approximately 50 hours.
During our hands-on session, using Dylan’s warhammer and scythe against enemies at close range immediately created a different combat rhythm from the original Control. Instead of relying primarily on ranged attacks, the new system encourages players to enter the danger zone and use Aberrant’s transformations to control encounters.
Accessibility Remains Part of the Design #
For players who are less comfortable with demanding action games, Control: Resonant retains extensive assist options.
Players can enable features such as invincibility and one-hit kills directly through the settings. This allows the difficulty to be tailored toward players who primarily want to experience the story and atmosphere.
The result is an interesting contrast: the same surreal environments that create tension can become considerably less threatening when players enable powerful assist features.
Because story details remain under embargo, the broader narrative and later gameplay sequences will have to wait until the relevant restrictions are lifted.
🌆 From the Oldest House to a Surreal Manhattan #
Moving from the Oldest House into Manhattan fundamentally changes the spatial structure of the game.
Remedy does not simply describe Control: Resonant as an open-world game. Instead, the city is structured as a large, open-ended collection of handcrafted districts, each with its own visual identity, optional encounters, side activities, and exploration opportunities.
The result is closer to a hub-based structure than a traditional massive open world.
The map exists primarily to support the narrative and provide meaningful exploration rather than simply filling a huge geographic area with activities. This approach gives Remedy more control over environmental storytelling while still providing considerably more freedom than the original game’s largely enclosed environments.
That philosophy is also reflected in the game’s rendering technology.
💡 Path Tracing Pushes Manhattan’s Lighting Further #
The original Control was already an important milestone for PC graphics.
At launch, it was among the early games to combine real-time ray tracing with DLSS and supported multiple ray-traced effects, including reflections, transparent reflections, diffuse lighting, contact shadows, and debris effects.
Control: Resonant continues that technical lineage with a significantly more advanced rendering stack built around DLSS 4.5.
The game supports technologies including:
- Path Tracing
- Ray Reconstruction
- DLSS Super Resolution
- Dynamic Multi Frame Generation
- DLAA
Among these technologies, Path Tracing represents the most significant change to the game’s lighting model.
Traditional rasterization approximates many lighting interactions, while conventional ray tracing typically applies ray calculations selectively to individual effects. Path Tracing takes a more unified approach by tracing light through the scene and accounting for multiple interactions and bounces.
That is particularly valuable in Control: Resonant.
Manhattan contains dense architecture, reflective surfaces, dramatic artificial lighting, dark interiors, supernatural effects, and constantly changing environments. A unified lighting model allows these elements to interact more consistently.
Path-Traced Reflections #
Screen Space Reflections (SSR) are limited by what the camera can currently see. If an object or light source exists outside the screen, SSR cannot directly use that information to produce a reflection.
Path-Traced Reflections remove much of this limitation by tracing rays through the environment.
The difference becomes especially obvious on vehicle surfaces and windows. Reflections can contain buildings, streets, and other environmental details that are outside the camera’s immediate view.
This makes reflective materials feel more physically connected to the surrounding world rather than simply reflecting whatever happens to be visible on screen.
Path-Traced Indoor Lighting #
Indoor environments benefit from Path Tracing in a different way.
Light can bounce from walls, floors, furniture, and other surfaces, creating more natural indirect illumination and contact shadows.
Without sufficient indirect lighting, objects can appear visually disconnected from their surroundings, producing the familiar “floating” effect associated with simplified lighting models.
With Path Tracing enabled, furniture such as office chairs can receive more believable illumination from surrounding surfaces, while shadows and light transitions become more coherent.
Path-Traced Shadows #
Path-Traced Shadows also account for the characteristics of individual light sources.
Large light sources can produce softer penumbras, while smaller sources generate sharper shadow boundaries. The resulting shadows better reflect the apparent size, direction, and position of the light source.
In Control: Resonant, this becomes particularly noticeable around architectural structures, barricades, and other objects exposed to Manhattan’s artificial and supernatural lighting.
🧠 DLSS 4.5 Ray Reconstruction Rebuilds the Image #
Path Tracing dramatically increases rendering complexity.
A GPU cannot practically trace an enormous number of rays for every pixel in every frame, especially at high resolutions. Instead, real-time Path Tracing works with limited samples, which creates noisy intermediate images that must be reconstructed into a stable final frame.
This is where DLSS 4.5 Ray Reconstruction becomes particularly important.
Ray Reconstruction uses an AI model trained on NVIDIA supercomputers to replace traditional hand-tuned denoising approaches with neural reconstruction. Rather than simply smoothing the noisy output, the system analyzes available spatial and temporal information to reconstruct a higher-quality image.
DLSS 4.5 further integrates denoising and Super Resolution into a more unified AI rendering pipeline. The system can analyze information across individual frames as well as between successive frames, improving detail, stability, and consistency.
The technology is particularly valuable in a game such as Control: Resonant, where lighting conditions and object placement can change rapidly.
The Second-Generation Transformer Model #
DLSS 4.5 builds on NVIDIA’s second-generation Transformer architecture.
The Transformer-based approach allows the rendering system to process a broader set of image information when reconstructing the final frame. In practical terms, this can improve stability during camera movement, changing illumination, and complex reflections.
During our testing, enabling Ray Reconstruction produced visible differences in several difficult rendering situations.
Mirror reflections became more accurate, translucent glass surfaces revealed additional reflection information, and diffuse reflections on walls became more apparent.
These are not simply brightness adjustments. They demonstrate how neural reconstruction can recover visual information that would otherwise be difficult to maintain when the underlying Path-Traced image contains limited samples.
🎮 DLSS Super Resolution and DLAA Complete the Pipeline #
Ray Reconstruction is only one component of the DLSS 4.5 stack.
DLSS Super Resolution renders the scene internally at a lower resolution and then uses AI reconstruction to produce the target output resolution.
This approach is particularly important for Path Tracing because lighting calculations can become extremely expensive at native 4K resolution. Rendering fewer pixels internally reduces the raw workload while allowing AI reconstruction to restore much of the detail required for a high-resolution final image.
For systems with sufficient performance headroom, DLAA takes a different approach.
Instead of reducing the internal rendering resolution, DLAA focuses on AI-based anti-aliasing at native resolution. This makes it more suitable for configurations where maximum image quality is prioritized over additional performance.
Together, Super Resolution and DLAA give players different ways to balance image quality and rendering performance.
🚀 Dynamic Multi Frame Generation Pushes Frame Rates Higher #
DLSS 4.5 also expands the role of Frame Generation.
Dynamic Multi Frame Generation can generate multiple additional frames between traditionally rendered frames. With the maximum 6x multiplier, the system can generate up to five additional frames.
The dynamic component is particularly interesting because the multiplier does not necessarily have to remain fixed.
Players can define a target frame rate, allowing the system to adjust frame generation according to the current workload. This can help avoid generating unnecessary frames when the GPU is already producing enough frames to exceed the display’s effective refresh rate.
The objective is therefore not simply to maximize a benchmark number. It is to balance frame rate, system latency, and available GPU resources according to the actual workload.
📊 GeForce RTX 50 Series Performance #
To examine the upper limits of the rendering pipeline, we pushed the game’s graphics settings and Path Tracing to maximum and tested different GeForce RTX 50 series configurations.
On the GeForce RTX 5080, enabling DLSS 4.5 Dynamic Multi Frame Generation at the 6x setting with Performance mode allowed the game to exceed 220 FPS at 4K with maximum graphics and Path Tracing.
Without Multi Frame Generation, performance remained below 80 FPS under the same conditions.
System latency increased from approximately 30 ms to around 50 ms, but the resulting frame-rate increase was substantial.
At 1440p, the RTX 5080 exceeded 270 FPS with the same DLSS 4.5 configuration, while the GeForce RTX 5070 Ti surpassed 230 FPS.
At 1080p, the GeForce RTX 5060 Ti 16GB exceeded 200 FPS with maximum graphics, Path Tracing, and DLSS 4.5 Dynamic Multi Frame Generation enabled in Performance mode.
These figures illustrate how dramatically AI-assisted rendering changes the performance envelope of a heavily path-traced title.
🔬 From Ray Tracing Showcase to Neural Rendering Showcase #
The technological progression from Control to Control: Resonant is particularly interesting.
The original game arrived at a time when real-time ray tracing was still relatively new to consumer GPUs. DLSS helped compensate for the substantial performance cost of enabling advanced ray-traced effects.
Seven years later, Control: Resonant represents a different stage of the same evolution.
Path Tracing provides a more unified lighting model, while DLSS 4.5 uses neural reconstruction to make that computationally expensive pipeline more practical.
Ray Reconstruction is therefore not merely another performance feature. It directly affects image quality by helping reconstruct reflections, indirect lighting, and other difficult-to-sample visual information.
Meanwhile, Super Resolution reduces the number of pixels that need to be rendered natively, and Dynamic Multi Frame Generation increases perceived frame rates by generating additional frames.
The result is a rendering pipeline in which AI participates at multiple stages rather than simply acting as a final image upscaler.
🏁 Final Thoughts #
From Control to Control: Resonant, Remedy’s games continue to intersect with major developments in real-time graphics.
The original Control helped demonstrate the potential of real-time ray tracing and early-generation DLSS. Control: Resonant takes that concept much further by combining Path Tracing with DLSS 4.5 Ray Reconstruction, Super Resolution, and Dynamic Multi Frame Generation.
The most interesting upgrade is arguably Ray Reconstruction.
Path Tracing provides the underlying lighting model, but limited ray samples create a fundamental reconstruction problem. DLSS 4.5 addresses that problem with neural rendering, allowing the game to recover more stable and detailed lighting information without requiring brute-force ray counts.
At the same time, Super Resolution reduces the cost of rendering those complex scenes, while Multi Frame Generation expands the achievable frame-rate range.
Our hands-on testing suggests that this combination can push even the GeForce RTX 5060 Ti 16GB beyond 200 FPS at 1080p with maximum graphics and Path Tracing when the full DLSS 4.5 stack is enabled. Higher-end GPUs can push substantially further at 1440p and 4K.
More importantly, the technology changes more than the performance counter.
In a game built around distorted architecture, supernatural phenomena, unpredictable lighting, and surreal reflections, rendering consistency is part of the atmosphere itself. DLSS 4.5 Ray Reconstruction helps preserve that consistency while Path Tracing supplies a more coherent physical lighting model.
That makes Control: Resonant an especially interesting showcase of where modern PC graphics are heading: not simply toward more rays, but toward a rendering pipeline where traditional graphics techniques and neural reconstruction work together to create increasingly complex real-time worlds.