Why does anything less than 16ms of a response time of a 60Hz monitor matter?

[Vozella]

60 frames per second is 60 Hz in a second. 

1 second equals 1000 milliseconds. 

1000/60=(50/3=16(2/3)=16.6...)

 

So, if every frame is 16 or 17 milliseconds at 60fps, why is it that people want monitors that have 5ms of response time or less? 

I'm asking becase I'm going for a 8ms monitor and I was worried at first but now it doesn't make sense. 

-

[144Hz would be

1000/144=(125/18=6(17/18)=6.94...)

-

1000/x=1

x=1000

 

You would need a thousand frames persond on a monitor for that 1ms to matter.]

-

I mean, am I doing something wrong here? If so, help me understand.

Edited June 26, 2015 by Vozella

[cragger89]

Marketing.

[Senzelian]

If you have a response time of 1 minute it takes one whole minute to actully get the image on the screen.

So you press a key and after one minute you see what happened. It does't matter how long the picture is visible.

See below what Glenwing wrote. Makes much more sense than my bullsh*t. 

[79wjd]

Let's not forget that there is no standardized way of measuring response times (well, there is, but not one that companies follow) -- i.e. response times are meaningless figures....just like Dynamic Contrast Ratios. 

[Glenwing]

How are latency and refresh rate related?

Good question, this is something I've been asked several times now. If a 60Hz monitor refreshes at 60Hz, then each frame is 16.667ms apart. Doesn't that mean there will be a minimum of 16.667ms of latency on a 60Hz monitor? Even if your monitor only takes, say, 2ms to process the input and have it ready to show, it will be another 14.667ms before the next frame appears on the screen anyway right?

Well, not exactly. Remember that we're talking about input latency here, the delay between receiving some input from the computer and actually showing it on the screen. The key thing to note here is that while there is a 16.667ms period between each frame on a 60Hz monitor, that's a 16.667ms waiting time since the last frame was shown, but your input could come at any time.

For example, let's say you have a 60Hz monitor, so the time between frames is 16.667ms, and this monitor takes exactly 2.000ms to process any input signal (the "real" latency). For the sake of simplicity let's also say there's zero latency between hitting a button on your mouse or keyboard and your GPU sending the monitor a new frame incorporating that input, so your monitor recieves the new information the instant you press a button.

If you press a button just after the monitor has refreshed, it will be another 16.667ms before you see it on the screen. So you will feel 16.667ms of latency rather than 2ms.

But on the other hand, if you press a button just 2.001ms before the next refresh, the display processes the signal in 2.000ms, and 0.001ms later the display refreshes, so you only feel 2.001ms of latency.

But on the other other hand, if you press a button 1.999ms before the next refresh, by the time the display has finished processing the input it's too late! The next refresh has already passed and you just missed it, the new frame can't appear until the next next refresh, in another 16.666ms. So the latency you would feel here is an entire refresh cycle (16.667ms) plus the additional 1.999ms from the previous cycle. (In the real world this would actually introduce tearing because the display wouldn't have finished drawing the frame so soon after a new refresh cycle started, but that's beside the point)

The point is that the measured latency actually fluctuates within a certain range. The latency from any given command depends on how your input happens to sync up with a monitor's refresh cycle. This is why any proper latency test must be conducted with many measurement trials, not just one or a few. The best-case scenario is always the monitor's actual processing latency, you'll never measure less latency than that. The worst case scenario is always the processing latency plus the frame time, in this case 2.000ms + 16.667ms. You can never measure latency worse than that unless of course there is interference from other factors (i.e. your peripherals or your computer or software is introducing an additional delay between your button press and sending an updated image to the monitor).

In our example, the latency for any individual button press can be anything between 2.000ms and 18.667ms (2.000ms+16.667ms) and the exact value will be different on every keypress. Monitors with a higher refresh rate will of course have a tighter latency range (8.333ms for 120Hz displays, 6.944ms for 144Hz displays) and monitors with lower refresh rates will have a larger latency range. On a 30Hz monitor, maximum latency for a keystroke is at least 33.333ms no matter how fast the monitor processes the signal.

[DildorTheDecent]

e-peen m8. 

[Glenwing]

Also, response time has nothing to do with latency, but I assume you actually meant latency.

[Vozella]

Also, response time has nothing to do with latency, but I assume you actually meant latency.

Isn't response time the one that when it's too high things get blurry? People call it either motion blur or ghosting.

I'm talking about that one. The time a pixel takes to change color.

[NeatSquidYT]

Isn't response time the one that when it's too high things get blurry? People call it either motion blur or ghosting.

I'm talking about that one. The time a pixel takes to change color.

I'm not sure, but apparently it is noticable even at 60Hz. Don't ask me how. My 7ms response monitor has 0 ghosting for me.

[manikyath]

Isn't response time the one that when it's too high things get blurry? People call it either motion blur or ghosting.

I'm talking about that one. The time a pixel takes to change color.

latency is why my tv is unusable for gaming, response time is why my tv looks like a washed out picture.

 

YAY free stuff.

[Glenwing]

Isn't response time the one that when it's too high things get blurry? People call it either motion blur or ghosting.

I'm talking about that one. The time a pixel takes to change color.

 

Yes, that's response time. That's independent of the refresh rate also, because the frame isn't constantly changing. Just because there's 16.667ms between each frame doesn't mean it's changing for the entire 16.667ms. It only takes, say 2ms, or 5ms to change, and then holds that image for the rest of the frame time. So the less time it takes, the less blur there is. But response time numbers don't really represent response time anyway, so don't worry about it too much