Skip to main content
numora.
numora.

Frame Time to FPS Calculator: Convert Frame Time to FPS Instantly

Convert frame time (ms) to frames per second and back β€” instantly see your GPU's real-world performance

EverydayByΒ Numora Everyday TeamPublished byΒ Numora Editorial Team

Try the calculator

Math runs in your browser. No sign-up.

Assumptions
ms
FPS
ms
FPS from frame time
β€”

FPS and frame time averages hide per-frame variance. Always review the full frame-time graph from a capture tool for a complete picture of smoothness.

Embed this calculator on your own site β€” drop in a one-line iframe.
Try a scenario
Quick takeaway

**Frame time and FPS are exact reciprocals: FPS = 1000 Γ· frame_time_ms.** At 60 FPS every frame must render in exactly 16.667 ms; at 144 FPS the budget shrinks to 6.944 ms. The number most players watch β€” average FPS β€” hides spikes: a game averaging 100 FPS but spiking to 50 ms on 1% of frames will stutter visibly even though the average looks fine. Always check your 1% low alongside the average.

What is a frame time to fps?

Frame time and frames per second are exact mathematical reciprocals: FPS = 1000 / frame_time_ms, where 1000 is the number of milliseconds in one second by SI definition. At 60 FPS every frame must render in 16.667 ms; miss that budget and the monitor repeats the previous frame, producing judder. Average FPS hides this β€” the 1% low frame time (99th-percentile frame time) is the metric that exposes stutter. This calculator converts in both directions and computes your 1% low FPS and stutter ratio from a single frame-time capture. The formula and all constants are sourced to the BIPM SI Brochure (9th edition, 2019).

The formula

FPS = 1000 / frame_time_ms | frame_time_ms = 1000 / FPS
  • FPS β€” frames rendered per second
  • t_ms β€” time to render one frame, in milliseconds
  • 1000 β€” milliseconds per second (exact SI definition: 1 s = 1000 ms)

Source: Reciprocal unit conversion: FPS = 1000 / frame_time_ms, derived from the SI definition 1 s = 1000 ms.

Worked examples

160 FPS target β€” standard 60 Hz monitor

Inputs
fps: 60frameTimeMs: 16.667lowOnePercentFrameTimeMs: 25
Walkthrough

The most common gaming target is 60 FPS. Applying FPS = 1000 / t_ms: 1000 / 16.667 = 59.998 FPS β‰ˆ 60 FPS (the tiny error is rounding 16.Μ„6Μ„ to 16.667). The inverse confirms it: t_ms = 1000 / 60 = 16.6Μ„ ms per frame. With a 1% low frame time of 25.0 ms the 1% low FPS = 1000 / 25.0 = 40 FPS. The stutter ratio = 25.0 / 16.667 = 1.50Γ—, sitting right on the boundary between smooth and mild variance. A player on a 60 Hz monitor would notice occasional brief dips below 40 FPS, though the average frame time looks acceptable.

2144 Hz high-refresh β€” competitive gaming

Inputs
fps: 144frameTimeMs: 6.944lowOnePercentFrameTimeMs: 10
Walkthrough

For a 144 Hz monitor the per-frame budget is t_ms = 1000 / 144 = 6.9Μ„4 ms. Rounding to three decimal places gives 6.944 ms, so FPS = 1000 / 6.944 = 144.01 FPS (rounding artifact only). With a 1% low frame time of 10.0 ms: 1% low FPS = 1000 / 10.0 = 100 FPS. Stutter ratio = 10.0 / 6.944 = 1.44Γ—, comfortably below 1.5Γ—. This is a smooth profile β€” the worst 1% of frames still deliver 100 FPS, well above the 60 FPS threshold most players consider the floor for competitive play.

330 FPS locked β€” console-style target

Inputs
fps: 30frameTimeMs: 33.333lowOnePercentFrameTimeMs: 50
Walkthrough

Many console and cinematic-style PC games target 30 FPS. The frame budget is t_ms = 1000 / 30 = 33.3Μ„ ms. Entering 33.333 gives FPS = 1000 / 33.333 = 30.00 FPS. A 1% low frame time of 50.0 ms corresponds to 1% low FPS = 1000 / 50.0 = 20 FPS and a stutter ratio of 50.0 / 33.333 = 1.50Γ—. Even at a locked-30 target, spikes to 50 ms are perceptible as judder β€” the 33 ms frame budget leaves almost no headroom before frames cross into the next refresh cycle on a 30 Hz output.

How to use this calculator

  1. Frame time β€” The time in milliseconds your GPU takes to render one frame. Find this in tools like FrameView, MSI Afterburner/RTSS, or CapFrameX.
  2. Frames per second (FPS) β€” The average number of frames rendered per second. Leave blank if you are converting from frame time above.
  3. 1% low frame time β€” The 99th-percentile frame time (the slowest 1% of frames). Report from CapFrameX, FCAT-OC, or FrameView. Leave blank if unknown.
  4. Read the result. Use the worked examples below to sanity-check against a known scenario.

Common mistakes and edge cases

Treating Average FPS as the Full Story

Average FPS is the arithmetic mean of all frames in a session. If 999 frames take 7 ms and one frame takes 700 ms, the average is about 7.7 ms (β‰ˆ 130 FPS) β€” but that 700 ms spike is a 0.7-second freeze that any player would notice. Always pair the average with the 1% low or a full frame-time graph. Tools like CapFrameX export percentile data alongside averages.

Rounding Frame Time Before Dividing

Entering 16.7 ms instead of 16.667 ms gives 1000 / 16.7 = 59.88 FPS instead of exactly 60 FPS. When diagnosing whether a game is hitting a locked frame-rate target, use the full precision value your capture tool reports. The formula has no built-in error; rounding the input is the only source of discrepancy.

Confusing 1% Low FPS with Minimum FPS

The 1% low is a percentile, not the absolute minimum. A single catastrophic 2000 ms stall raises the 0.1% low but may barely move the 1% low if it happened only once in an hour-long session. Minimum FPS is the worst single frame and is dominated by outliers. The 1% low is statistically more robust and is the industry-standard stutter metric for this reason.

Assuming a Higher Average FPS Means a Higher Refresh Rate Helps

A monitor renders at its native refresh rate regardless of GPU output. If your GPU delivers 80 FPS to a 144 Hz monitor, 64 of the 144 refresh cycles per second will repeat the previous frame. The monitor does not create new frames β€” it only displays what the GPU provides. A higher refresh rate improves perceived smoothness only when the GPU can sustain the corresponding frame time budget.

How small changes affect your result

Because FPS = 1000 / t_ms, the relationship is non-linear: a 1 ms improvement at low frame times (high FPS) is worth far more than a 1 ms improvement at high frame times (low FPS).

At 60 FPS (16.667 ms baseline)
** - Saving 1 ms (β†’ 15.667 ms) gains 3.8 FPS (β†’ 63.8 FPS) - Saving 5 ms (β†’ 11.667 ms) gains 25.7 FPS (β†’ 85.7 FPS) - Adding 5 ms overhead (β†’ 21.667 ms) loses 13.8 FPS (β†’ 46.2 FPS)
At 144 FPS (6.944 ms baseline)
** - Saving 1 ms (β†’ 5.944 ms) gains 23.3 FPS (β†’ 168.3 FPS) - Adding 1 ms overhead (β†’ 7.944 ms) loses 18.3 FPS (β†’ 125.9 FPS)
At 240 FPS (4.167 ms baseline)
** - Saving 1 ms (β†’ 3.167 ms) gains 56.7 FPS (β†’ 315.7 FPS) - Adding 1 ms overhead (β†’ 5.167 ms) loses 46.7 FPS (β†’ 193.5 FPS)
The non-linearity is why CPU bottlenecks matter more at high frame rates
a 1 ms CPU overhead that is negligible at 60 FPS costs nearly 57 FPS at 240 FPS. This is the mathematical basis for the advice to minimise background CPU load in competitive gaming.

Frame time budgets at common refresh-rate targets

Target FPSMonitor HzFrame time budget (ms)1% low target (ms)Stutter headroom (ms)
30 FPS30 Hz33.33350.00016.667
60 FPS60 Hz16.66722.2225.556
120 FPS120 Hz8.33311.1112.778
144 FPS144 Hz6.9449.2592.315
165 FPS165 Hz6.0618.0812.020
240 FPS240 Hz4.1675.5561.389
360 FPS360 Hz2.7783.7040.926

1% low target assumes a stutter ratio of 1.5Γ— (the smooth threshold). Stutter headroom = 1% low target βˆ’ average frame time budget. All values computed from FPS = 1000 / t_ms.

Frequently asked questions

What is the formula to convert frame time to FPS?
FPS = 1000 / frame_time_ms. Because there are exactly 1000 milliseconds in one second (BIPM SI Brochure, 9th ed., 2019), dividing 1000 by the number of milliseconds per frame gives frames per second. For example, 16.667 ms per frame equals 1000 / 16.667 = 59.998 FPS β‰ˆ 60 FPS.
How do I convert FPS back to frame time in milliseconds?
frame_time_ms = 1000 / FPS. At 60 FPS: 1000 / 60 = 16.667 ms. At 144 FPS: 1000 / 144 = 6.944 ms. At 240 FPS: 1000 / 240 = 4.167 ms. The relationship is a simple reciprocal β€” both directions use the same constant (1000 ms/s).
Why is frame time better than FPS for measuring smoothness?
FPS is an average that hides individual spikes. A game averaging 100 FPS could have one frame taking 100 ms β€” a full tenth of a second freeze β€” while all others take 7 ms, and the average still looks fine. Frame time captures every individual frame, so a histogram or 1% low metric reveals stutters that FPS averages conceal.
What is the 1% low FPS and why does it matter?
The 1% low is the frame rate at the 99th-percentile frame time β€” the slowest 1% of frames in a benchmark run. It is the standard stutter metric because it represents realistic worst-case performance during gameplay, not an outlier spike. If your average is 144 FPS but your 1% low is 40 FPS, most players will notice judder even though the average looks great.
What frame time do I need for 60 FPS?
Exactly 16.6Μ„ ms (16.667 ms rounded to three decimal places). Any frame that takes longer than 16.667 ms on a 60 Hz display either delays the next frame or causes a dropped frame, which the monitor repeats the previous frame to fill.
What frame time do I need for 144 FPS?
1000 / 144 = 6.944 ms per frame. Your GPU must complete every frame within that budget for the monitor to display each one at the correct time. A single 10 ms frame at 144 Hz is already 1.44Γ— over budget and will cause one frame to be delayed by roughly half a refresh cycle.
What is a good stutter ratio?
A stutter ratio (1% low frame time Γ· average frame time) below 1.5Γ— is generally considered smooth. Between 1.5Γ— and 2.0Γ— produces occasional perceptible hitches. Above 2.0Γ— produces noticeable, distracting stutter that most players will want to investigate β€” common causes include background CPU load, shader compilation stutters, or VRAM pressure.
How do I measure frame time on my PC?
Three free tools capture per-frame timing: CapFrameX (open source, exports CSVs with P99 and P1 metrics), MSI Afterburner with RTSS (overlay + logging), and NVIDIA FrameView (all GPU vendors, exports frame-time CSVs). For the most accurate measurement β€” capturing what the monitor actually receives rather than what the API reports β€” use FCAT-OC methodology.

Frame Time to FPS glossary

Frame time
The elapsed wall-clock time, measured in milliseconds, that the GPU spends rendering a single frame. It is the direct measurement; FPS is derived from it.
Also called: frame duration Β· render time per frame
FPS (frames per second)
The number of frames rendered in one second, equal to 1000 divided by the average frame time in milliseconds. FPS is a derived metric β€” not a direct measurement.
Also called: frame rate Β· framerate
1% low
The FPS or frame-time value at the 99th percentile of a captured frame-time distribution: 99% of frames were faster than this value. It quantifies worst-case stutter without being skewed by a single spike.
Also called: P99 frame time Β· 99th-percentile frame time Β· 0.1% low (more extreme variant)
Stutter ratio
The ratio of the 1% low frame time to the average frame time. A ratio above 2Γ— means the worst frames take more than twice as long as average, which most players perceive as a noticeable hitch.
Also called: frame-time variance ratio
Refresh rate
How many times per second a monitor redraws its display, measured in Hz. A frame time longer than one refresh interval (e.g., >16.67 ms on a 60 Hz monitor) causes a dropped or repeated frame.
Also called: Hz Β· display frequency
FCAT-OC
Frame Capture Analysis Tool β€” Overlay Capture, a methodology developed by NVIDIA for capturing frame-time data at the monitor's output signal rather than at the API level, providing the most accurate real-world measurement.
Also called: FCAT

How we built this calculator

Methodology

The relationship between frame time and FPS is a strict mathematical reciprocal derived from the SI definition of the second. One second contains exactly 1000 milliseconds (BIPM SI Brochure, 9th ed., 2019, Table 6). Therefore:

Formula source
Reciprocal unit conversion: FPS = 1000 / frame_time_ms, derived from the SI definition 1 s = 1000 ms
Last updated
2026-04-23
Calculation runs
Client-side only
NE
WRITTEN BY
Numora Everyday Team
NE
PUBLISHED BY
Numora Editorial Team

Reviewed on 2026-04-23 Β· Next review: 2027-04-23

See editorial policy

Sources & references

Every numeric assumption traces to a primary source.

  1. BIPM SI Brochure, 9th edition (2019) β€” Table 6: non-SI units accepted for use with the SI (millisecond definition)INT
  2. NVIDIA FrameView β€” frame-time capture and analysis tool documentationUSA
  3. CapFrameX β€” open-source frame-time analysis tool (percentile methodology documentation)INT
  4. NVIDIA FCAT-OC methodology β€” frame-capture analysis at display outputUSA
  5. VESA DisplayHDR and Adaptive Sync specifications β€” refresh rate and frame-pacing definitionsINT
  6. Numora Editorial Policy. numora.net/editorial-policy