Stutter, Frametime & Glossary: Understanding 1% Lows, Frame Pacing & Hitches

Tier	Hardware Requirements
MINIMUM SPEC	N/A (measurement tools/hardware)
RECOMMENDED SPEC	N/A (measurement tools/hardware)

Average FPS tells only part of the story. Two systems with identical average frame rates can deliver vastly different experiences - one buttery smooth, the other plagued by frustrating stutters and hitches. To understand how a game feels in motion, we need to look beyond simple averages to the underlying frame timing data that reveals consistency and stability. This glossary explains the frametime metrics used to evaluate and optimize gaming performance.

Core Concepts: Frametime vs FPS

Frametime measures how long it takes to render a single frame (in milliseconds), while FPS counts frames per second. They’re inversely related: FPS = 1000 / frametime. Lower frametimes mean higher FPS and smoother gameplay. When analyzing performance, frametime data is more valuable than FPS because it’s linear and additive, making statistical analysis more meaningful.

Key Performance Metrics Explained

1% Lows (and 0.1% Lows)

The 1% low FPS is commonly reported as the average frame rate of the worst 1% of frames rendered during a benchmark period. It shows how bad performance gets during the most demanding moments, but the exact calculation varies by tool.

What it measures: Performance during the worst 1% of gameplay moments Why it matters: Reveals sustained performance dips that impact perceived smoothness How to interpret:

Higher 1% low FPS = More consistent performance (better)
A 1% low FPS close to your average FPS indicates minimal variation
Example: Average FPS 100 with 1% low 60 means significant drops in worst 1% of frames

Note: Some tools report 1% low as frametime (99th percentile of frametimes), others as FPS. Some implementations average the lowest 1% of FPS values, while others use a time-weighted threshold. Always check units and method.

Frame Pacing

Frame pacing refers to consistency of time intervals between successive frames. Perfect frame pacing means each frame arrives at exactly the same interval (e.g., every 16.67ms for 60 FPS).

What it measures: Variability in frame delivery timing Why it matters: Poor frame pacing causes micro-stutter that makes gameplay feel uneven even at high average FPS How it’s measured: Standard deviation of frametimes, frame time variance, percentile-based frametime analysis (e.g., 95th percentile), frame time delta graphs

Visual indicators:

Consistent frametime graph = Good frame pacing
Spiky frametime graph = Poor frame pacing (stutter)
Regular patterns = Specific pacing issues (e.g., alternating slow/fast frames)

Hitches

Hitches are noticeable pauses or stutters that disrupt immersion and control responsiveness.

What they measure: Sudden, significant frametime increases creating perceptible delays Why they matter: Directly impact gameplay feel, causing missed inputs, visual jumps, frustration How they’re identified:

Frametime spikes well above average (often several times the normal value)
Corresponding drops in instantaneous FPS
Visible stutter/pause during gameplay playback
Often linked to asset loading, shader compilation, background processes

Perceptibility thresholds:

Hitches become more obvious as their duration grows. Spikes around 50ms or longer are easy for most players to notice, while shorter spikes may still bother sensitive or competitive players. These values correspond to the frame times of 20 FPS, 30 FPS, and 60 FPS and are commonly used as reference points rather than fixed rules.

Additional Important Metrics

0.1% Lows

Shows the average worst 0.1% of frames - revealing the absolute bottom of performance most likely to cause severe stuttering.

Average FPS

Arithmetic mean of all frame rates during benchmark. Useful as headline but hides inconsistencies that 1% and 0.1% lows reveal.

99th Percentile Frametime

Frametime below which 99% of frames fall. Directly relates to 1% low FPS when converted.

Frame Time Spike Ratio

Ratio of maximum frametime to average frametime - higher values indicate more severe outliers.

Stutter Frequency

Rate at which frametime exceeds a threshold (e.g., >50ms to render).

Measurement Tools: CapFrameX & PresentMon

PresentMon: The Foundation

Intel’s GameTechDev team developed PresentMon as the underlying ETW (Event Tracing for Windows) tool capturing high-performance graphics API data with minimal overhead.

Key capabilities:

Low-overhead ETW tracing for DirectX (including DXGI), OpenGL, and Vulkan
Captures CPU, GPU, display frame durations and latencies
Tracks dropped frames and presentation delays
Provides hardware telemetry (power, temperature, utilization) via vendor APIs
Command-line interface with extensive configuration options
Outputs detailed CSV data for further analysis
Designed for minimal overhead

What it captures:

CPUFrameTime / MsCPUBusy: CPU time preparing each frame
GPUFrameTime / MsGPUTime, MsGPUBusy: GPU time rendering each frame
DisplayLatency / MsUntilDisplayed: Time from present until frame is displayed
GPUBusy: Percentage of time GPU is actively rendering
And dozens of other hardware/software metrics

CapFrameX: The Analysis Frontend

CapFrameX builds on PresentMon’s data collection to provide an accessible, feature-rich benchmarking/analysis environment tailored for gaming performance evaluation.

Key features:

Intuitive GUI for configuration and analysis
Real-time overlay during gameplay (via RTSS integration)
Automated run history/aggregation for statistical significance
Outlier detection/handling to improve data quality
Multiple visualization methods (frametime graphs, FPS charts, L-shape, etc.)
Percentile-based metric calculation (1% lows, 0.1% lows, etc.)
Sensor monitoring (CPU/GPU temps, clocks, power, utilization)
CSV export for external analysis (Excel, Python)
Support for multiple graphics APIs (DirectX, OpenGL, Vulkan) on Windows
Minimal setup with sensible defaults

Analysis views:

Analysis Tab: Detailed individual capture examination with frametime graphs, FPS charts, metric summaries
Comparison Tab: Side-by-side evaluation of multiple runs to compare settings, hardware, driver versions
Overlay: Real-time display of key metrics during gameplay (FPS, frametime, 1% lows, etc.)

Practical Application: What These Metrics Reveal

Diagnosing Performance Issues

Different performance problems show characteristic patterns in frametime data:

GPU Bottleneck:

Consistently high GPUFrameTime
GPU utilization near 100%
CPU utilization significantly lower than GPU
Frametime increases correlate with graphical settings changes

CPU Bottleneck:

Consistently high CPUFrameTime
CPU utilization near 100%
GPU utilization significantly lower than CPU
Frametime increases correlate with CPU-intensive settings (draw distance, physics, NPC count)

Memory Bottleneck:

Stuttering correlated with texture loading or asset streaming
High memory utilization with frequent page faults
Performance improves with reduced texture quality or resolution

Storage Bottleneck:

Periodic hitches every few seconds during open world traversal
Correlated with disk usage spikes
Improved with faster storage (SSD vs HDD) or defragmentation

Driver/Software Issues:

Irregular frametime patterns not correlated with gameplay events
Spikes at regular intervals (suggesting background tasks)
Improvement after driver reinstallation or software updates

Optimizing for Smoothness

Understanding these metrics enables targeted optimization:

To improve 1% lows:

Reduce graphical settings causing worst performance spikes
Enable frame rate caps to prevent excessive GPU load during simple scenes
Optimize background processes and startup applications
Consider CPU/GPU upgrades if consistently bottlenecked

To improve frame pacing:

Enable V-Sync or adaptive sync (FreeSync/G-Sync) to match refresh rate
Use frame rate limiters to prevent GPU overwork
Ensure adequate power delivery and cooling to prevent throttling
Update drivers and check for game-specific optimization guides

To eliminate hitches:

Identify and disable background processes causing regular spikes
Move games to faster storage (NVMe SSD)
Increase system RAM to reduce paging
Update drivers and check for known issues with specific hardware combinations
Consider increasing virtual memory page file size

Metrics and Perceived Smoothness Relationship

The relationship between frametime metrics and perceived smoothness is usually described in qualitative terms rather than fixed scientific thresholds:

Excellent Smoothness:

1% low FPS close to average FPS (within roughly 10%)
Frame time standard deviation very low
No noticeable hitches in typical gameplay
Consistent frametime graph with minimal variation

Good Smoothness:

1% low FPS within roughly 20% of average FPS
Frame time standard deviation low
Occasional minor hitches (rarely noticeable)
Mostly consistent frametime with minor spikes

Noticeable Stutter:

1% low FPS well below average (e.g., within 30-50%)
Frame time standard deviation moderate
Frequent noticeable hitches
Clear frametime inconsistency patterns

Poor Smoothness:

1% low FPS less than half of average FPS
Frame time standard deviation high
Regular disruptive hitches
Obviously uneven frametime delivery

These are rules of thumb; individual sensitivity, display technology, and game genre all affect how stutter is perceived.

Best Practices for Accurate Measurement

To ensure frametime metrics reflect true gameplay performance:

Warm up system: Typically run the game 10-15 minutes before benchmarking for thermal equilibrium and shader caching
Consistent test segments: Benchmark the exact same gameplay section each time (specific level, boss fight, traversal route)
Control background: Close unnecessary apps, disable startup items, pause updates during testing
Manage thermals: Ensure adequate cooling, verify stable temperatures during testing
Adequate sample size: Typically capture 2-3 minutes or more per run, and consider averaging multiple runs
Disable interfering overlays: Turn off non-essential overlays (Discord, Xbox Game Bar, etc.) that add noise
Check drivers: Use clean driver installations, verify no known bugs
Consider monitor limits: Remember display refresh rate and response time also impact perceived smoothness

The Bottom Line

While average FPS is commonly cited, it’s frametime-derived statistics like 1% lows, frame pacing measurements, and hitch frequency that determine how smooth a game feels in motion. A system with modest 60 FPS average but excellent consistency often feels smoother than erratic 100 FPS average with frequent stutters.

Tools like PresentMon and CapFrameX have democratized access to this performance analysis level, allowing players, reviewers, and developers to move beyond superficial benchmarks to understand and optimize actual gameplay experience. By learning to interpret these metrics, you gain power to diagnose issues, validate optimizations, and make informed hardware decisions based on how games actually feel to play - not just what average FPS counters say.

As games push graphical boundaries and hardware grows more complex, understanding these frametime fundamentals remains essential for anyone serious about gaming performance - whether optimizing competitive setups, reviewing AAA releases, or simply getting more enjoyment from existing hardware.

Stutter, Frametime & Glossary: Understanding 1% Lows, Frame Pacing & Hitches