Frame Limiters Benchmarked: Real-Life Frame Pacing Comparison

Why "60 FPS" Can Still Feel Stuttery

Most players evaluate performance using a simple in-game FPS counter. If the counter reads "60", they assume the game is running smoothly.

However, the FPS counter only counts how many total pictures your PC made during a full one-second window. It tells you nothing about when those frames arrived.

• Perfect Frame Pacing (Smooth): At 60 FPS, exactly one frame arrives every 16.67 milliseconds like clockwork. Motion looks fluid and responsive.
• Uneven Frame Pacing (Stutter): One frame arrives in 4ms, the next takes 30ms, and another takes 8ms. You still got 60 frames in that second, but the irregular rhythm creates noticeable stutter.

Test Methodology & Benchmark Setup

Accurate benchmarking requires an independent referee. External frame limiters work by hooking into the game's rendering pipeline. If you rely on a limiter's own built-in overlay to measure performance, it only sees frame delivery from its own internal perspective, which will be biased and hide stutters that happen further down the display line.

Furthermore, running multiple limiting tools at the same time causes them to compete on the exact same hook chain, interfering with each other’s timing and distorting benchmark results.

To ensure 100% fair, accurate, and unbiased results:

• Unbiased External Capture: We used CapFrameX (ETW-based) analyzing MsDisplayBetweenFrames. Instead of hooking inside the game or merely measuring when the engine renders a frame, this tracks the exact millisecond duration each frame physically stays on your display.
• One Limiter at a Time: Every frame limiter was tested strictly in isolation.
• Clean Game Restarts: The game was fully closed and restarted after every single test run to guarantee no residual hooks, injected DLLs, or memory conflicts remained.

We evaluated 6 different frame limiting configurations across 3 key targets (30 FPS, 60 FPS, and 61 FPS):

• In-Game Limiter: The built-in framerate limiter inside the game engine settings.
• Driver-Based Limiter (NVIDIA App / NVCP): The global "Max Frame Rate" cap configured in GPU driver software.
• RTSS Async: The default asynchronous limiter mode in RivaTuner Statistics Server.
• RTSS Front Edge (AW - Active Wait): RivaTuner's spin-wait limiter.
• RTSS Front Edge (PW - Passive Wait): RivaTuner's sleep-wait limiter.
• Special K (Normal): Special K's normal limiter.
• framepacer v0.4: framepacer's limiter.

Test 1: 30 FPS - The Console Stability Test

At a 30 FPS target, each frame must stay on screen for exactly 33.33 milliseconds. This test demonstrates how each limiter handles heavy single-player pacing where consistency is critical.

30 FPS Benchmark Results

Analysis & Findings

The data clearly separates the tools into two distinct tiers:

• High Precision Tier (framepacer, RTSS Front Edge, Special K): These tools delivered a practically flawless 99.8% to 99.9% consistency rate. Consecutive frames stayed within 2ms of the 33.33ms target. The 0.2% lows stayed above 29 FPS, delivering the smooth feel typically associated with dedicated game consoles.
• Fluctuating Tier (In-Game & RTSS Async): The in-game limiter struggled significantly. Almost 40% of its frames had timing errors greater than 2ms, with some frames jumping by more than 8ms to 12ms. This variance dragged 0.2% lows down to 25.6 FPS, causing visible hitching despite a solid 30 FPS average.

Test 2: 60 FPS - The Standard PC Gaming Test

At 60 FPS, the target frametime drops to 16.67 milliseconds. Because the timing window is cut in half, any delay or inaccuracy in the limiter becomes twice as noticeable to your eyes.

60 FPS Benchmark Results

Analysis & Findings

The 60 FPS test exposes a critical flaw in built-in game limiters:

• The In-Game Limiter Collapse: Only 24% of frames were delivered within 2ms of ideal timing. Over 75% of frames drifted by 4ms to 12ms. This causes the game to rapidly alternate between fast frames (12ms) and slow frames (24ms), producing persistent micro-stutter even though the overlay shows a constant 60 FPS.
• RTSS Async Degradation: In Async mode, only 42% of frames maintained tight < 2ms variance, with 0.2% lows dipping below 50 FPS.
• framepacer & Front Edge Stability: framepacer and RTSS Front Edge held 99.5% of all frames in the ultra-tight < 2ms bracket, maintaining a smooth frametime graph and strong 0.2% lows.

Test 3: 61 FPS - Odd Number Test

Gamers frequently set odd FPS caps (such as 61 FPS, 59 FPS, or 141 FPS on VRR displays) to stay within their VRR range, avoid input lag of V-Sync, or prevent LFC behavior.

We benchmarked the official NVIDIA App Driver Limiter against external tools at an odd 61 FPS limit to test driver-level pacing accuracy and non-standard cadence handling.

61 FPS Benchmark Results

Analysis & Findings

The 61 FPS test highlights how different limiters handle non-standard frame intervals:

• Special K, framepacer & RTSS Front Edge: These tools delivered near-perfect pacing, holding 99.1% to 99.4% of all frames within the ultra-tight < 2ms variance bracket. Their 0.2% lows remained consistently strong above 55–56 FPS.
• RTSS Async Drift: In Async mode, RTSS managed only 39% of frames under < 2ms variance, with 0.2% lows dipping to 50.7 FPS.
• The Driver Limiter Collapse: The NVIDIA App driver-based limiter showed severe frame pacing degradation. Only 26% of frames were within 2ms of the target, causing 1% lows to drop to 41.9 FPS and 0.2% lows to plunge to 34.7 FPS (a massive 43% drop below average FPS). Over 74% of frames suffered from noticeable jitter.

Why Do Limiters Perform So Differently?

Each limiter is built with a different goal in mind. Some are made for maximum smoothness, while others are made for minimum input lag:

Summary: Which Limiter Should You Use?

If your goal is eliminating stutters and achieving flat frametimes, relying on in-game or driver-level limiters is not enough. Using a high precision frame limiter like framepacer ensures that every single frame is delivered with exact millisecond precision to your screen.