DeviceBenchFree Hz to ms converter
Hz to ms, and the frame time behind it
Type a rate and the period comes back in milliseconds; type a period and the rate comes back, because 1000 divided by either one produces the other. So 60 Hz gives each frame 16.67 ms to happen in, 144 Hz gives it 6.94 and a monitor advertised at 165 Hz gives it 6.06 — and the table below carries eighteen rates with the milliseconds each one takes off the row beneath it, which is where an upgrade stops being worth the money. The number you type is taken at face value: this page divides, it does not measure, so it cannot tell you whether your display is running at the rate printed on its box.
- 100% free
- No signup
- Both directions
- 18 rates listed
- Two decimals
Jump to a rate
1.50 frames at 60 Hz — round that up, because a picture only changes on a boundary.
Frame time
16.67ms
1000 ÷ 60 = 16.6667 ms
| Rate | Frame time | Gained on the row above | Where you meet it |
|---|---|---|---|
| 41.71 | — | Film transferred to NTSC video — 24 × 1000/1001 | |
| 41.67 | 0.04 | Cinema projection and most film-look camera modes | |
| 40.00 | 1.67 | PAL broadcast frame rate | |
| 33.33 | 6.67 | Phone video by default, and the cap on plenty of console modes | |
| 20.83 | 12.50 | The bottom of a typical adaptive-sync range | |
| 20.00 | 0.83 | PAL television field rate, still standard in Europe | |
| 16.68 | 3.32 | What a television calling itself 60 Hz is usually doing | |
| 16.67 | 0.02 | Every laptop panel and office monitor out of the box | |
| 13.33 | 3.33 | The first step up most budget 1080p monitors offer | |
| 11.11 | 2.22 | The floor most standalone VR headsets hold | |
| 10.00 | 1.11 | Ultrawides, and 50 Hz television doubled | |
| 8.33 | 1.67 | Phones, tablets and current consoles in performance mode | |
| 6.94 | 1.39 | The mainstream gaming monitor for a decade | |
| 6.06 | 0.88 | 144 Hz panels sold with the factory overclock applied | |
| 4.17 | 1.89 | Esports panels, TN and fast IPS | |
| 2.78 | 1.39 | Competitive-tier LCDs | |
| 2.08 | 0.69 | The top of current OLED desktop panels | |
| 1.85 | 0.23 | The fastest desktop refresh rate sold |
This page divides, it does not measure — the rate you type is taken at face value, including the one printed on a monitor box that the operating system was never actually told to use. The figure it hands back is one link of a chain that also contains the mouse, the game engine, the compositor, the cable and the panel’s own pixel response, none of which shrink because the frame budget did.
How to convert a refresh rate into frame time
One division, then the two figures underneath it that the division does not say out loud.
Put a number in either field
Both boxes are live and each one rewrites the other, so a rate off a monitor listing goes in the top field and a period off a benchmark log goes in the bottom one. The ten buttons under them jump straight to the rates people actually run, and every row of the table below is clickable for the same reason.
Read the two lines below the headline before you quote it
Twice the period is what one blown deadline costs: the picture that should have changed does not, so the same image is held for two slots and the next one arrives that much later. Half the period is the other figure worth knowing, because an event arriving at a random moment waits on average half a frame for the display to have any opportunity to show it.
Compare rows using the gain column, not the Hz
The third column of the table is milliseconds removed from the row above, and it is the honest way to weigh an upgrade: the jump from 60 Hz to 144 Hz takes 9.72 ms out of every frame, while 480 Hz to 540 Hz takes 0.23. Rates are a ratio and frame times are a subtraction, which is why doubling a big number stops feeling like anything.
Technical specifications
| The identity | milliseconds = 1000 ÷ hertz, and hertz = 1000 ÷ milliseconds; both fields are live and editing either one rewrites the other |
|---|---|
| Displayed precision | Two decimals on the headline, four on the printed working and a whole-microsecond figure beside it — enough to keep 500 Hz at 2.00 ms distinct from 540 Hz at 1.85 ms |
| Reference table | 18 rates from 23.976 Hz to 540 Hz, each with its period and the milliseconds it removes from the row above |
| Anchor values | 60 Hz → 16.67 ms, 75 Hz → 13.33, 120 Hz → 8.33, 144 Hz → 6.94, 165 Hz → 6.06, 240 Hz → 4.17, 540 Hz → 1.85 |
| Broadcast fractions included | 59.94 Hz → 16.68 ms and 23.976 Hz → 41.71 ms, the 1000/1001 rates NTSC video still carries |
| Derived figures | Double the period for one missed deadline, half of it for the average wait an event faces, and frames per minute for reading a session log |
| Accepted range | 0.1 Hz to 10,000 Hz in either direction, so a mouse polling at 8,000 Hz and a 24-second time-lapse interval both convert |
| What leaves the tab | Nothing — the conversion is a single division with no request behind it, and no rate is remembered between visits |
Frequently asked questions
Why is 60 Hz written as 16.67 ms and not 16.6 or 17?
Because the true value is 16.666… ms and it never terminates, so every written form is a rounding of it. Two decimals is the useful cut: it keeps 500 Hz (2.00 ms) apart from 540 Hz (1.85 ms) and it stops the error mattering when you multiply. If you round to 17 ms and count a minute of frames you are 20 ms out; the calculator carries the full value internally and only rounds for display, which is why the microsecond line reads 16,667 rather than 17,000.
What is 59.94 Hz, and why does my television use it instead of 60?
It is 60 multiplied by 1000/1001, a fraction inherited from the moment color was added to black-and-white NTSC broadcasting and the frame rate had to be nudged so the color subcarrier would not beat against the sound carrier. Its period is 16.68 ms against 60 Hz's 16.67 — a difference of 17 microseconds that sounds like nothing and is not: a source running at exactly 60 and a display locked to 59.94 drift a whole frame apart every 16.7 seconds, which is the judder people notice on video that is otherwise fine. The same fraction gives 23.976 Hz, at 41.71 ms, which is how 24 fps film is carried on NTSC video.
My monitor advertises 1 ms. Is that its frame time?
No, and the two are not even measuring the same thing. The number on the box is pixel response, usually gray-to-gray, meaning how long a subpixel takes to finish changing shade once it has been told to; the frame time here is how often the panel is willing to be told anything at all. A 240 Hz screen gives each frame 4.17 ms, so a genuine 1 ms response fits comfortably inside one frame and a 5 ms response does not — that is the comparison the two numbers are for, rather than one being a better version of the other.
How much does a single dropped frame actually cost?
Exactly one extra period, which the converter shows as the doubled figure. At 60 Hz a missed deadline means 33.33 ms pass between one picture and the next instead of 16.67, and at 144 Hz it means 13.89 ms instead of 6.94. That is why a stutter on a fast panel is shorter but easier to spot: the eye is comparing it against a stream of very short frames, so the interruption stands out even though it lasts a fifth as long.
Does vertical sync change these numbers?
It does not change the period, but it forces everything to land on multiples of it. With vsync on, a frame that arrives even a fraction late waits for the next whole slot, so on a 60 Hz display the sustainable rates are 60, then 30, then 20 — 16.67 ms, 33.33 ms, 50 ms — and there is nothing in between. That staircase is why frame rate under vsync appears to halve rather than sag, and why adaptive sync, which lets the panel move its own boundary, removes the effect instead of smoothing it.
Can I use this for mouse polling or an audio buffer?
Yes — the arithmetic is the period of any repeating event, not a display-specific formula. A 1,000 Hz mouse reports every 1.00 ms and an 8,000 Hz one every 0.125 ms, a 125 Hz office mouse every 8.00 ms; a 48 kHz audio stream puts 0.0208 ms between samples. The fields accept anything from 0.1 Hz to 10,000 Hz, which covers a time-lapse interval at one end and the fastest pointing device sold at the other.
Does it work with no connection, and does anything I type get sent?
It works offline and nothing is sent, because the entire tool is one division that runs in this tab. There is no request behind a conversion, no rate is written to storage, and the fields reopen on 60 Hz however many times you have used them — which also means a result you want to keep needs the copy button rather than a bookmark.
About frame time, fractional rates and why the gains shrink
A rate and a period are the same fact told from opposite ends, and the conversion is a reciprocal: one second holds 1000 milliseconds, so a rate of n per second puts 1000 ÷ n milliseconds between events. What makes the millisecond form the more useful one is that it is a budget rather than a score. At 60 Hz everything the machine has to do to produce the next picture — read the input, run the simulation, issue the draw calls, let the compositor assemble it — has to fit inside 16.67 ms; at 144 Hz the same work has 6.94 ms; at 240 Hz, 4.17 ms. Frame rate hides that squeeze behind a growing number, which is why engine profilers and driver overlays quote frame time instead, and why a graph of milliseconds shows a hitch that an averaged frames-per-second counter smooths away entirely. Measuring the budget your own tab is hitting is a different job, and it belongs to the FPS test, while the rate the panel is genuinely accepting is what the refresh rate test times.
Two of the rows in the table are not whole numbers, and they are the reason video files behave strangely more often than displays do. When color was added to NTSC television the frame rate was multiplied by 1000/1001 to keep the color subcarrier from interfering with the sound, and the fractions survived into digital video: 59.94 Hz and 23.976 Hz are still what a great deal of footage is authored at. Their periods, 16.68 ms and 41.71 ms, sit a hair away from the whole-number versions, and a hair is enough. Play a 23.976 file on a display locked to exactly 24 and a frame has to be repeated or dropped roughly every 42 seconds; run 60 Hz content against a 59.94 Hz panel and the two slip a full frame apart every 16.7 seconds. Neither is a fault in the file or the screen, and no amount of adjusting the picture settings will touch it — the fix is matching the output rate to the source, which is exactly the kind of decision this conversion exists to support.
The part people are usually really asking about is whether the next monitor is worth it, and frame time answers that better than hertz does because the saving is a subtraction. Going from 60 Hz to 144 Hz removes 9.72 ms from every frame. Going from 144 Hz to 240 Hz removes 2.78. From 240 Hz to 360 Hz it is 1.39, and from 480 Hz to 540 Hz, 0.23 — one part in forty of what the first upgrade gave, for a step that costs considerably more. Where that saving sits in the whole delay between a hand and a photon is the subject of the input lag test, which times the press-to-painted-frame span the frame boundary quantizes. The rest of the settings that decide whether the budget is met at all are on their own pages: a wider view puts more geometry inside the same frame, which is what the FOV calculator is really trading, and carrying an aim across games so a faster panel does not also change how far your hand has to travel is what the sensitivity converter handles. If the frames are also being captured rather than merely displayed, doubling the rate roughly multiplies the data your encoder has to move, and the bitrate calculator prices that in gigabytes and upload speed.
What the converter keeps
Every number on this page is worked out by JavaScript running in the tab you are reading it in. Nothing you type, paste or open is uploaded, logged or kept, which is also why the tools carry on working after you disconnect from the network.
A conversion here is one division against a number you typed, so there is nothing to store and nothing to send: no rate is written to local storage, no preset is remembered, and the pair of fields reopens on 60 Hz whatever you last worked out. The Copy button writes to your clipboard and stops there.