FPS vs Milliseconds: What Frame Time Is and How to Convert It

Fps (frames per second) and milliseconds per frame are two ways of saying the same thing. You convert with one division: fps = 1000 / ms and ms = 1000 / fps, because one second has 1000 milliseconds. 60 fps equals 16.67 ms per frame; 144 fps equals 6.94 ms; 30 fps equals 33.33 ms.

Fps is what the counters show, and milliseconds are what smoothness is made of. Understanding the difference explains why two PCs with the same average fps can feel very different, and it is the foundation for diagnosing any performance problem.

The formula and a conversion table

Each frame takes a slice of the second. The shorter the slice, the more frames fit. That slice is the frame’s time budget: everything the PC and the game have to do (simulate, prepare, draw) must fit inside it if you want to hold that rate.

Frames per second	Milliseconds per frame	Where it is used
30	33.33	Playable minimum; consoles and modest PCs
40	25.00	120 Hz modes that split the rate
60	16.67	Usual target on 60 Hz displays
75	13.33	Office displays with a slightly higher rate
90	11.11	Virtual reality and mid-range displays
100	10.00	Comfortable point for almost everything
120	8.33	High-refresh displays
144	6.94	Standard gaming monitor
165	6.06	Common 144 Hz variant
240	4.17	Competitive games with a very powerful PC

To convert any figure there is the fps and milliseconds calculator, with a symptom and cause table. The table also tells you how much you have to save. Going from 60 to 90 fps does not need 30 more fps of “power” but trimming 5.56 ms from every frame. Going from 30 to 60 trims 16.67 ms. That is why 5 extra fps hurts when you start at 20 and barely registers when you start at 200: in milliseconds, 20 to 25 fps saves 10 ms, and 200 to 205 fps saves 0.12 ms.

Frame time: what you feel

A counter that says “100 fps” is an average of what happened in the last second. Frame time is the real time of each individual frame, and that is where you see what the eye perceives.

Imagine a second in which 99 frames take 8 ms each and one takes 100 ms. The total is 99 × 8 + 100 = 892 ms, so the counter will show about 112 fps, an excellent number. But that 100 ms frame is a pause of a tenth of a second, and it looks like a hitch. The average did not count it because it diluted it among 99 others.

Smoothness depends on two things:

• The rate: how many frames per second, that is, the average time per frame.
• The regularity: that frames arrive at similar intervals. This is sometimes called frame pacing.

A rate of 60 fps with very regular 16.67 ms frames looks better than an average of 90 fps with 40 ms spikes. So when you compare PCs or settings, look at the frame time graph, not only the number.

Averages and the 1 % and 0.1 % lows

To summarize the spikes without reading an entire graph, we use lows (1 % lows and 0.1 % lows):

Metric	What it summarizes	What it tells you
Average	All frames equally	General performance, spikes excluded
1 % low	The slowest 1 % of frames in the capture	The habitual hitches
0.1 % low	The slowest 0.1 %	The strongest hitches, the ones you remember
Absolute minimum	The single slowest frame	One pause; useful to spot loading

Programs define these metrics with small differences (the average of the worst 1 %, or the 99th percentile of frame time), so always compare numbers computed with the same tool. Free tools such as PresentMon or CapFrameX calculate them from the time of every frame.

A practical rule: the closer the 1 % low is to the average, the more regular the game feels; if it is half the average or less, the hitches are obvious. It is not a standard but a starting point, and it depends on the genre and how demanding you are.

Three different “ms” when you play

In a conversation about performance, milliseconds appear in three contexts and they are easily mixed up:

What it is	What it measures	What improves it
Frame time	How long the PC takes to produce each frame	CPU, GPU, graphics settings
Input (system) latency	From the click or key press to the change on screen	More fps, low-latency technologies such as NVIDIA Reflex, a fast display
Ping (network latency)	Round trip of a message to a server	Wired connection, route, distance to the server

Raising fps lowers frame time and usually lowers input latency, but it does not touch ping. In an MMO it is worth knowing which of the three bothers you before you spend money. The third is covered in ping, latency and servers in an MMO, and low-latency technology in DLSS, frame generation and Reflex explained.

Which target to set

• Look at your display. A 60 Hz display shows at most 60 frames per second; the extra ones are discarded or tear the image. With 144 Hz, the natural target is to get close to 144.
• Prefer steady to high. Setting a cap a little below what your PC sustains gives more regular frames than leaving it free. How to do it with V-Sync and G-Sync or FreeSync is in V-Sync, G-Sync, FreeSync and frame limits.
• In online games with many entities, prioritize the lows. The busiest scenes are where frames drop, and that is the experience you will remember. When and why that happens is in why MMOs and survival games depend on the CPU.

How we read the performance of a game in development

COLONIZED is a space survival MMO in development (closed alpha applications are open and the first wave is planned for winter 2026-2027; it cannot be played yet). We measure the cost of each effect in milliseconds rather than fps, precisely because milliseconds add up. In our development build, at 3440×1440 on a GeForce RTX 5070 Ti, the volumetric clouds went from about 15.5 ms to 3.5 ms per frame. In fps that improvement would be confusing; in milliseconds you subtract and it makes sense. The full story is in our guide to optimizing Unreal Engine 5 games. These are development measurements, not system requirements or an fps promise.

If you want to measure your own PC, the practical guide is how to measure your PC’s performance. And if you want to know why your card is not working at full tilt, start with what a bottleneck is.