Measuring performance with the Frame Analyzer
Date Published
An overview of the Frame Analyzer
The Frame Analyzer is a tool within the IQON App designed to provide detailed insights into the performance characteristics of your game sessions. Under the hood, the IQON App leverages PresentMon, a utility developed by Intel, to reliably capture frame-level performance data.
Users can record gameplay sessions of variable capture lengths, enabling both quick checks and deeper analysis. The tool also includes features for comparing recordings and exporting data, making it well-suited for advanced debugging and visualization workflows.
Frame time vs Frames per second (FPS)
These two terms get used interchangeably in gaming, but they describe different aspects of performance.
FPS (frames per second) tells you how many frames are shown every second.
Frame time tells you how long each individual frame takes to appear on screen, usually measured in milliseconds.
They’re closely related: if frames take longer to render, you’ll see fewer of them each second. If they’re quick to render, FPS goes up. So while they describe the same system, they’re looking at it from different angles.
The key difference is what they emphasize:
- FPS shows overall speed — how much work is being done over time
- Frame time shows consistency — how evenly that work is delivered
And consistency turns out to matter more than people expect. Human perception is very sensitive to irregular timing. A steady stream of frames feels smooth, even if it’s not especially fast.
For example, a locked 30 FPS (where every frame arrives at a steady rhythm) can feel smoother than a 60 FPS experience that occasionally stutters. Those small interruptions break the illusion of motion.
The illusion of FPS
If there’s one takeaway, it’s this: FPS can hide problems that frame time makes obvious.
Imagine this sequence:
- Most frames arrive quickly and evenly
- One frame suddenly takes much longer than the others
An FPS counter will average everything together and still report a high number. On paper, performance looks great.
But what you actually feel is that one delayed frame.
From the player’s perspective, here’s what happens:
- The game shows one frame
- The next frame takes too long to render
- The display keeps showing the old frame while it waits
The result is a brief pause — a hitch or stutter.
Frame time makes this visible immediately as a spike. FPS smooths it out into an average, which is why it often misses the issue.
What metrics are available?
Now that we understand why frame time matters, you can rest easy knowing that the IQON App shows both FPS and frame time statistics for your game recordings. The table below highlights a few of the most important metrics, with short explanations for both frame time and FPS to help you understand what each reveals about your gameplay performance.
IQON also tracks additional statistics beyond what’s shown here, giving you a deeper view into performance and helping advanced users analyze hitches, stutters, and other detailed frame behaviors.
Metric | Fps | Frametime |
|---|---|---|
Avg | Average FPS → throughput, but hides spikes | Average frame time in ms → overall smoothness |
P99 | 1% lowest FPS → the “1% low” FPS most noticeable to players | 99th percentile frame time → 1% of frames are slowest → shows stutter |
1% avg | Average of lowest 1% FPS → shows the 1% low FPS more clearly than raw P1 | Average of slowest 1% frames → captures stutter without extreme outliers |
Min/Max | Highest / lowest FPS → frames per second extremes | Fastest / slowest frame time → best/worst smoothness |
Tips for using the tool
To get meaningful results from the Frame Analyzer, consistency is key. Small changes in what’s happening on screen can have a big impact on performance, so it’s important to control what you can.
Measure in the same scenario
Try to record in a repeatable section of the game — the same area, camera angle, and general activity. Comparing two runs only makes sense if the workload is similar. For example, standing in a busy town or hub area will produce very different results compared to active gameplay.
Avoid “noisy” situations
Highly variable scenes (crowded hubs, dynamic events, background downloads, etc.) can introduce randomness into your data. These are useful for stress testing, but not ideal for clean comparisons.
Record for long enough
Very short captures can miss intermittent issues. A slightly longer recording increases the chance of catching spikes or stutters and gives a more representative view of performance.
Focus on changes, not absolute numbers
The tool is most powerful when comparing runs — before vs after a setting change, driver update, or hardware tweak. Look for differences in consistency (frame time spikes, variance), not just average FPS.
Use frame time to investigate issues
If something feels off in-game, frame time is usually where the answer is. Spikes and irregular patterns will often point directly to the cause of stutter, even when FPS looks fine.
Other features
In addition to the above, the Frame Analyzer also gives you a look at how your various sensors performed during the session:
Finally, the tool also lets you compare different game sessions; perfect for before and after optimization checks:
