Frame time vs FPS: why a high average can still feel choppy
Read frame-time spikes, distinguish repeatable stutter from low FPS and compare one route without hiding slow frames.
In brief
FPS counts frames over a period; frame time measures how long an individual frame takes. A good average can conceal occasional long frames that you feel as a hitch.
Looking at the time graph can show whether a settings change reduces stutter rather than merely raising the average.
Read the graph in milliseconds
For illustration, smooth 16.7 ms frames correspond to about 60 FPS, while one 80 ms frame can cause a visible hitch. These are arithmetic examples, not measurements of your game. Check the scale: a tall graph line does not mean the same delay in every overlay.
Repeat the part that stutters
Mark the location or action and run it again after the first load. If the initial spike disappears, first-run loading may be involved. If it recurs in the same crowded scene, examine scene complexity and CPU activity. If it appears randomly during downloads or recording, compare with that background work paused.
Report the result honestly
Keep the run length, route, hardware and settings consistent. If a tool reports 1% lows, use the same tool and method for comparisons; definitions can vary. Do not present a short, different scene as proof of a universal performance gain.
Before you act
A single spike does not identify its cause. Loading, shader work, CPU delays and storage can all contribute; measurement tools also have their own overhead.
What to do next
- Enable the game’s frame-time display or a trusted performance overlay you already use.
- Repeat a short route with the same resolution and settings, noting where spikes occur.
- Change one suspected factor and compare both average FPS and long-frame behaviour.
