Ray Tracing Explained: How Real-Time Lighting Works

Ray tracing is a rendering technique that simulates the paths of individual light rays to produce physically accurate reflections, shadows, and lighting in real time. Unlike the traditional rasterization method that approximates lighting, ray tracing traces rays from the camera into the scene and follows how they bounce off surfaces, calculating illumination from the geometry of light itself.

Nvidia introduced real-time ray tracing to consumer graphics with its RTX cards, and AMD and Intel followed with their own hardware. This guide defines ray tracing, contrasts it with rasterization, describes what it improves, explains the dedicated hardware that accelerates it, quantifies its performance cost, covers the upscaling that offsets that cost, and introduces path tracing as its advanced form.

What Is Ray Tracing?

Ray tracing is a rendering technique that simulates the travel of light rays through a scene to calculate reflections, shadows, and global illumination with physical accuracy. The technique casts rays from the virtual camera into the three-dimensional scene, determines which surfaces each ray strikes, and then traces secondary rays toward light sources and reflective surfaces to compute color and brightness. Because it models how light actually propagates and bounces, ray tracing produces accurate reflections, soft shadows, and indirect lighting that older methods approximate or omit.

Ray tracing existed for decades in offline film rendering, where a single frame could take hours, but real-time ray tracing in games became practical only with dedicated GPU hardware. Nvidia, AMD, and Intel now build ray tracing acceleration into their graphics cards, enabling the technique to run within the frame time budget of interactive applications. The article on how graphics cards render explains the broader pipeline this technique sits within.

How Does Ray Tracing Differ From Rasterization?

Ray tracing differs from rasterization because rasterization projects geometry onto the screen and approximates lighting, while ray tracing simulates the actual paths of light rays to compute lighting directly. Rasterization, the long-standing real-time method, converts triangles into pixels and applies lighting through techniques such as shadow maps and screen-space reflections, which are fast but produce artifacts and cannot represent light that originates off-screen. Ray tracing instead follows rays from the camera, computing how light reaches each point from its true source, including reflections of objects not visible on screen.

Rasterization is far faster and remains the foundation of real-time rendering, so current games combine the two: rasterization draws the base image, and ray tracing adds specific effects such as reflections or shadows. This hybrid approach delivers improved lighting accuracy while keeping the frame rate playable. Pure ray tracing of an entire scene remains more demanding than most hardware can sustain at high frame rates.

What Does Ray Tracing Improve in Games?

Ray tracing improves several lighting effects that rasterization approximates poorly:

• Reflections become accurate because ray tracing reflects the true scene, including objects and light sources outside the camera view, which screen-space reflections cannot show.
• Shadows gain correct softness and contact accuracy because traced rays determine exactly which light reaches each surface, producing penumbra that shadow maps approximate.
• Global illumination reproduces indirect light bouncing between surfaces, so a colored wall casts its color onto nearby objects, which rasterization cannot compute directly.
• Ambient occlusion darkens crevices and contact points accurately by tracing how much ambient light reaches each point, replacing the screen-space approximation.

These effects combine to produce lighting that responds correctly to scene geometry and off-screen elements. Reflections show the full environment rather than only what the camera sees, shadows soften with distance from the occluder, and indirect light fills a scene the way real light does.

The visual gain is most apparent in scenes with reflective surfaces, complex lighting, and large indoor or outdoor environments. The cost of computing these effects in real time is why dedicated hardware acceleration is required, and why upscaling is commonly paired with ray tracing to keep frame rates playable.

What Hardware Accelerates Ray Tracing?

Ray tracing is accelerated by dedicated hardware units, called RT cores on Nvidia cards and Ray Accelerators on AMD cards, that compute ray-geometry intersections far faster than general shaders. The most expensive step in ray tracing is testing where each ray intersects the scene geometry, which is organized into an acceleration structure called a bounding volume hierarchy. Nvidia RT cores, introduced with the RTX architecture, perform these intersection tests in fixed-function hardware, freeing the general-purpose shader cores for other work.

AMD Ray Accelerators, built into its RDNA architecture, perform the equivalent function, and Intel Arc includes ray tracing units in its Xe architecture. Each generation has increased the throughput of these units, narrowing the performance gap between ray-traced and rasterized rendering.

Without dedicated acceleration, computing ray intersections on general shaders would be too slow for real-time frame rates. The presence and generation of this hardware is a major factor when selecting the best graphics card for gaming with ray tracing enabled.

What Is the Performance Cost of Ray Tracing?

The performance cost of ray tracing is a frame rate reduction commonly between 30 and 60 percent at high settings, because tracing rays and computing their intersections demands significant additional GPU work. Enabling ray-traced reflections, shadows, and global illumination adds millions of ray calculations per frame on top of the rasterized base image. The exact cost depends on the number of effects enabled, their quality settings, the resolution, and the capability of the ray tracing hardware.

A card with stronger RT cores or Ray Accelerators loses a smaller share of its frame rate than a weaker card. At 4K, where the pixel count is highest, the absolute cost is greatest.

This performance penalty is the central trade-off of ray tracing: it improves visual accuracy at a substantial frame rate cost, which is why upscaling and frame generation are commonly enabled alongside it to restore playable frame rates. The cost has fallen across hardware generations but remains significant.

How Do Upscaling and Frame Generation Offset the Cost?

Upscaling and frame generation offset ray tracing’s cost because upscaling renders the frame at a lower internal resolution and reconstructs it to the target, while frame generation inserts interpolated frames to raise the displayed frame rate. Nvidia DLSS, AMD FSR, and Intel XeSS render internally at a lower resolution, such as 1440p for a 4K target, then use algorithms, including AI models in DLSS and XeSS, to reconstruct a sharp full-resolution image. This recovers much of the frame rate lost to ray tracing at a small image-quality cost.

Frame generation goes further by synthesizing entirely new frames between rendered ones, raising the displayed frame rate, though it adds latency. Together, these techniques make ray tracing playable at high resolutions: a card that loses half its frame rate to ray tracing can recover most of it through upscaling.

The pairing of ray tracing with upscaling is now standard in games that support it. Comparing the Nvidia and AMD approaches shows differences in upscaling image quality and ray tracing throughput.

What Is Path Tracing?

Path tracing is the advanced form of ray tracing that simulates the full propagation of light through a scene, tracing many bounces per ray to compute complete global illumination. Where typical game ray tracing applies traced effects selectively on top of a rasterized image, path tracing models the entire lighting of a scene by following rays through multiple bounces between surfaces and light sources. This produces the most physically accurate result, with unified reflections, shadows, and indirect lighting computed together rather than as separate effects.

Path tracing is the method used in film rendering, and recent games have introduced it as an optional high-end mode, sometimes labeled full ray tracing. The computational cost is far higher than selective ray tracing, so path tracing in real time depends heavily on upscaling and frame generation, and on the most capable flagship hardware. Path tracing represents the direction real-time rendering is moving as ray tracing hardware grows more powerful with each generation.

Which Games and Hardware Support Ray Tracing?

Ray tracing is supported by a wide range of current games and by every recent graphics card generation from Nvidia, AMD, and Intel. On the hardware side, Nvidia RTX cards from the RTX 20-series onward, AMD Radeon RX 6000-series and later, and Intel Arc cards all include ray tracing acceleration, exposed through the DirectX Raytracing (DXR) and Vulkan ray tracing APIs. On the software side, many major titles support ray-traced reflections, shadows, and global illumination, with some offering full path tracing as a high-end option.

The level of support varies: some games apply ray tracing to a single effect, while others apply it broadly. Performance scales with the capability of the card’s ray tracing hardware, so a flagship card runs more ray-traced effects at higher frame rates than an entry card. A player wanting ray tracing should confirm both that the game supports it and that the graphics card has hardware capable of running it at the target resolution.

Rasterization Versus Ray Tracing

The table below compares rasterization and ray tracing by their method, lighting accuracy, performance cost, and typical use in current games.

Key Takeaways

• Ray tracing simulates the paths of light rays to compute accurate reflections, shadows, and global illumination.
• Rasterization projects geometry and approximates lighting, so current games combine it with ray-traced effects.
• RT cores on Nvidia and Ray Accelerators on AMD compute ray-geometry intersections in dedicated hardware.
• The performance cost of ray tracing is commonly a 30 to 60 percent frame rate reduction at high settings.
• Upscaling (DLSS, FSR, XeSS) and frame generation recover much of the frame rate lost to ray tracing.
• Path tracing is the advanced form that traces full light propagation, demanding the most capable hardware and upscaling.

Last Thoughts on Ray Tracing

Ray tracing brings physically accurate lighting to real-time rendering by simulating the paths of light rays rather than approximating them as rasterization does. The technique improves reflections, shadows, global illumination, and ambient occlusion, computing light from the true geometry of the scene including off-screen elements. Dedicated hardware, RT cores on Nvidia and Ray Accelerators on AMD, accelerates the ray-intersection calculations that make real-time ray tracing possible.

The technique carries a substantial performance cost, commonly a 30 to 60 percent frame rate reduction, which upscaling and frame generation offset. Path tracing extends the method to full light propagation for the most accurate result.

As ray tracing hardware grows more capable each generation, the technique moves from an optional effect toward the standard lighting model. The computer hardware guide and the guide on graphics card architecture explain the silicon that executes it.