What Is Frame Generation?

Frame generation is a technology that inserts additional generated frames between the frames a graphics card renders, raising the displayed frame rate beyond what the card renders on its own. The generated frames are interpolated from neighboring rendered frames using motion vectors and, in current implementations, optical flow analysis, so the displayed image updates more often than the render rate alone allows. Frame generation differs from upscaling, which rebuilds each rendered frame at a higher resolution, and it differs from a true render frame rate, since the generated frames carry no new player input.

This article defines frame generation, describes the main implementations including DLSS 3 Frame Generation, FSR 3, and AMD Fluid Motion Frames, explains how generated frames are produced, covers the latency trade-off, identifies the best use cases, and describes the artifacts it can introduce. Frame generation improves perceived smoothness on a high-refresh display without lowering input lag. Each section answers one question about how frame generation works and when it helps.

What Is Frame Generation?

Frame generation is a technology that creates and inserts interpolated frames between rendered frames, raising the displayed frame rate above the rate at which the graphics card actually renders the scene. The graphics card renders a sequence of frames, and frame generation produces an additional frame between each pair, doubling or further raising the number of frames the display shows.

The generated frame is computed from the rendered frames around it rather than from the game engine. Frame generation has three defining traits:

• The frame insertion places a generated frame between two rendered frames, raising the count of frames the display shows each second.
• The interpolation source computes each generated frame from neighboring rendered frames and motion data, not from a new game-engine update.
• The displayed-rate increase follows from the inserted frames, so a 60 frames-per-second render can present as 120 displayed frames per second.

Frame generation raises the displayed frame rate, which is distinct from the rendered frame rate defined in the explanation of FPS in gaming. The technique is separate from the upscaling that lowers render cost, compared across vendors in the comparison of DLSS, FSR, and XeSS.

What Are DLSS 3, FSR 3, and AFMF?

DLSS 3 Frame Generation, FSR 3 Frame Generation, and AMD Fluid Motion Frames are the main frame generation implementations, each inserting generated frames between rendered ones, differing in hardware support and whether they operate inside the game or at the driver level. The three implementations share the core technique of frame insertion while differing in integration and compatibility. The implementations work as follows:

• DLSS 3 Frame Generation from Nvidia inserts generated frames using optical flow and motion data, running on supported RTX graphics cards, according to Nvidia’s documentation.
• FSR 3 Frame Generation from AMD inserts generated frames within a supported game and runs on a broad range of graphics cards, according to AMD’s documentation.
• AMD Fluid Motion Frames applies frame generation at the driver level, so it can work in games that do not integrate frame generation directly, according to AMD’s documentation.

In-game implementations such as DLSS 3 and FSR 3 use the game’s motion vectors, while a driver-level approach such as AMD Fluid Motion Frames works more broadly but without that engine data. The hardware support for each, and its pairing with upscaling, is compared in the comparison of DLSS, FSR, and XeSS.

How Does Frame Generation Work?

Frame generation works by analyzing two consecutive rendered frames, computing the motion between them with motion vectors and optical flow, and constructing an intermediate frame that represents the scene at a point in time between the two. According to Nvidia’s documentation, DLSS 3 uses an optical flow accelerator to estimate motion and generate the intermediate frame. The process follows three steps:

1. The motion estimation analyzes two rendered frames and the game’s motion vectors to determine how each object moves between them.
2. The frame construction builds an intermediate frame that places each object at its interpolated position partway between the two rendered frames.
3. The frame insertion displays the generated frame between the two rendered frames, raising the displayed frame rate.

Optical flow and motion vectors let the technology place moving objects accurately in the generated frame, though fast or complex motion is harder to interpolate. The motion data comes from the same rendering pipeline described in the explanation of how GPUs work, and the rendered frames that bound each generated frame are defined in the explanation of FPS in gaming.

Does Frame Generation Reduce Input Lag?

Frame generation does not reduce input lag, because the generated frames are interpolated from existing rendered frames and carry no new player input, so responsiveness is governed by the rendered frame rate rather than the higher displayed rate. The displayed frame rate rises, but the time between a player’s action and a frame that reflects it does not fall, and the buffering needed to interpolate can add a small amount of latency. The latency behavior has three parts:

• The unchanged input response means generated frames show no new input, so the game responds at the rendered frame rate, not the displayed one.
• The slight added latency can occur because the technology holds a rendered frame to interpolate the frame before it.
• The mitigation pairs frame generation with a low-latency mode such as Nvidia Reflex to offset the added delay, according to Nvidia’s documentation.

Frame generation improves smoothness without improving responsiveness, which separates a generated frame rate from a rendered one for competitive play. The latency it does not reduce is defined in the explanation of input lag in gaming, and methods to lower that delay appear in the guide to reducing input lag.

When Should You Use Frame Generation?

Frame generation is best used when the rendered frame rate is already high, the game is single-player or not latency-critical, and the display has a high refresh rate, since these conditions maximize the smoothness gain while minimizing the impact of the unchanged input lag. The technology adds the most value where displayed smoothness matters more than the lowest possible latency. The best use cases are listed below:

• An already-high base frame rate gives the cleanest result, since a high rendered rate means generated frames interpolate over small motion gaps with fewer artifacts.
• Single-player and visually driven games benefit most, since smoothness matters more than the lowest latency in these titles.
• A high-refresh monitor displays the added frames, so a 120-hertz or faster panel shows the smoothness gain that frame generation produces.

Frame generation is less suited to competitive games, where the rendered frame rate and low latency matter more than displayed smoothness, as the explanation of input lag in gaming details. A high base frame rate before generation is the target set in the guide to a good FPS for gaming.

What Artifacts Can Frame Generation Cause?

Frame generation can cause visual artifacts such as ghosting, warping around fast-moving objects, and errors at the user interface, because the interpolated frame is estimated rather than rendered and the estimate is imperfect during rapid or complex motion. The artifacts appear where motion estimation is hardest, and they are most noticeable at lower base frame rates. The common artifacts are listed below:

• Ghosting leaves a faint trail behind a moving object when the interpolation cannot place it cleanly between frames.
• Warping distorts edges around fast-moving objects where the estimated motion does not match the true path.
• Interface errors can appear on heads-up display elements that move differently from the scene, since they are not represented by world motion vectors.

Artifacts are reduced by a high base frame rate, since smaller motion gaps between rendered frames make interpolation more accurate. A low base frame rate produces both larger artifacts and a weaker latency profile, which ties the result back to the rendered frame rate defined in the explanation of FPS in gaming and the upscaling that can raise it in the comparison of DLSS, FSR, and XeSS.

What Is the Difference Between Frame Generation and a Native High Frame Rate?

A native high frame rate comes from the graphics card rendering every frame in full, so each frame carries new player input and reduces latency, while frame generation raises the displayed frame rate with interpolated frames that carry no new input and do not lower latency. Both raise the number of frames the display shows, but a native frame rate improves responsiveness as well as smoothness, whereas frame generation improves only smoothness. The two differ in three ways:

• The frame source differs because every native frame is rendered from a fresh game-engine update, while a generated frame is interpolated from neighboring rendered frames.
• The latency effect differs because a higher native frame rate lowers input lag, while frame generation leaves input lag at the rendered rate.
• The hardware demand differs because a native high frame rate requires a faster graphics card and processor, while frame generation raises the displayed rate without that rendering cost.

A native high frame rate is preferable where responsiveness matters, since it lowers the latency defined in the explanation of input lag in gaming, while frame generation suits cases where displayed smoothness is the goal. The rendered frame rate that sets the responsiveness baseline is defined in the explanation of FPS in gaming, and upscaling raises that rendered rate, as the comparison of DLSS, FSR, and XeSS covers.

Key Takeaways

• Frame generation inserts generated frames between rendered ones, raising the displayed frame rate above the render rate.
• DLSS 3, FSR 3, and AFMF are the main implementations, differing in hardware support and in-game versus driver-level operation.
• It works by interpolation, estimating motion between two rendered frames with motion vectors and optical flow.
• It does not reduce input lag, since generated frames carry no new input and can add slight latency.
• It is best with a high base frame rate, in single-player games, and on a high-refresh monitor.
• It can cause artifacts such as ghosting and warping, reduced by a higher base frame rate.

Last Thoughts on Frame Generation

Frame generation is a technology that inserts interpolated frames between rendered ones to raise the displayed frame rate, implemented in DLSS 3, FSR 3, and AMD Fluid Motion Frames through motion vectors and optical flow. It improves perceived smoothness without reducing input lag, works best at a high base frame rate in single-player games on a high-refresh display, and can introduce artifacts that a higher base rate reduces. Readers can continue with the comparison of DLSS, FSR, and XeSS, the explanation of input lag in gaming, the explanation of FPS in gaming, or the PC gaming guide hub for related concepts.