Input lag test program?
1 - 20 of 58 Posts
Joined
·
2,077 Posts
There isn't one. I don't think it will be possible to create a program to do that, as it would be reliant on knowing when you started a physical motion in relation to when it actually received the input.
Joined
·
1,213 Posts
You'll have to reverse engineer the thing. Input lag caused by smoothing takes place in the tracking algorithm.
From a theoretical standpoint: mlt04--> 9000fps ~400cpi out of a 22*22 array, thought for 125Hz: should have no correction algorithm whatsoever.
adns 3090--> 6400fps 1800/3500cpi out of a 30*30 array, though for 1000Hz: should have minimal correction (smoothing).
Now one thing I noticed is that high static friction combined with excessive weight give enough momentum to mask input lag a bit since you'll feel your hand stopping after the mouse really stops.
If after testing you can't feel a difference then the smoothing is low enough (opinion shared by many) on the 3500cpi 3090.
Always remember that computers are not real time
![smile.gif]()
From a theoretical standpoint: mlt04--> 9000fps ~400cpi out of a 22*22 array, thought for 125Hz: should have no correction algorithm whatsoever.
adns 3090--> 6400fps 1800/3500cpi out of a 30*30 array, though for 1000Hz: should have minimal correction (smoothing).
Now one thing I noticed is that high static friction combined with excessive weight give enough momentum to mask input lag a bit since you'll feel your hand stopping after the mouse really stops.
If after testing you can't feel a difference then the smoothing is low enough (opinion shared by many) on the 3500cpi 3090.
Always remember that computers are not real time
If you happen to have a camera capable of high framerates, you can record a short video where there is your mouse and your screen visible, then do a sharp knock on the mouse to move it, and then do the same thing with the other mouse, and then compare the video frame-by-frame to see the difference in which the image on the screen starts changing as opposed to the moment of the mouse moving. This obviously wont tell you the real input lag, but lets you compare two mouses to see fi there is a significant difference.
Joined
·
2,567 Posts
It also depends what you want to measure.
a) Mouse to PC
b) Mouse to Display
First choice is fully within responsibility of mouse manufacturer. Second one is affected by ... everything. Typical mouses are now USB only, but I still prefer PS/2 interface simply because mouse will receive its own interrupt (IRQ) and its not depending on CPU as USB mouses.
This is also why is measurement from blurbusters completely wrong if they wanted to measure Gsync parameters. The method measures difference between clicking the mouse and time when shooting appears on display. There is simply no reason to say that gsync monitor affects something when there is CPU, GPU, memory latency and absolutely everything in the way.
But yes, its true that you can use that method to measure two different mouses, but the results will me very inaccurate.
a) Mouse to PC
b) Mouse to Display
First choice is fully within responsibility of mouse manufacturer. Second one is affected by ... everything. Typical mouses are now USB only, but I still prefer PS/2 interface simply because mouse will receive its own interrupt (IRQ) and its not depending on CPU as USB mouses.
This is also why is measurement from blurbusters completely wrong if they wanted to measure Gsync parameters. The method measures difference between clicking the mouse and time when shooting appears on display. There is simply no reason to say that gsync monitor affects something when there is CPU, GPU, memory latency and absolutely everything in the way.
But yes, its true that you can use that method to measure two different mouses, but the results will me very inaccurate.
Joined
·
696 Posts
Just to point out something.
Input lag on a 125hz mouse is 8ms. Then there is a certain delay between the switch being pressed and the that signal being converted to a data signal. That is another 8ms or so (usually less). Total lag is 16 ms.
Now if you take the best mouse on the market you will get 1ms input lag and 1ms signal conversion. Total delay 2 ms.
Great 8 times less. Then we add a human. Which is supposed to use the mouse. And human input lag (or reaction time) is 220-250ms.
So whatever the mouse you have the biggest problem and the biggest lag is you.
Input lag on a 125hz mouse is 8ms. Then there is a certain delay between the switch being pressed and the that signal being converted to a data signal. That is another 8ms or so (usually less). Total lag is 16 ms.
Now if you take the best mouse on the market you will get 1ms input lag and 1ms signal conversion. Total delay 2 ms.
Great 8 times less. Then we add a human. Which is supposed to use the mouse. And human input lag (or reaction time) is 220-250ms.
So whatever the mouse you have the biggest problem and the biggest lag is you.
Quote:
reference mouse could be a IME 3.0 and then you could have a graph like in the mousebutton-latency comparision graph.
now someone with lots of mice, should get such highspeed camera and do it.
![biggrin.gif]()
this would be good enough to compare different mice.Originally Posted by Microx256
If you happen to have a camera capable of high framerates, you can record a short video where there is your mouse and your screen visible, then do a sharp knock on the mouse to move it, and then do the same thing with the other mouse, and then compare the video frame-by-frame to see the difference in which the image on the screen starts changing as opposed to the moment of the mouse moving. This obviously wont tell you the real input lag, but lets you compare two mouses to see fi there is a significant difference.
reference mouse could be a IME 3.0 and then you could have a graph like in the mousebutton-latency comparision graph.
now someone with lots of mice, should get such highspeed camera and do it.
Joined
·
2,269 Posts
Quote:
all the more reason to avoid even more "lag", or are you saying youd play the same with a 60hz monitor with 50ms+ gtg time, 60ms input lag, 30 fps ingame and a mouse with another 50ms or so input lag, why not play with 500ms ping while youre at it, i mean, you the human is the -only- limiting factor, right?
Wait, why even 60hz monitor, people can only see 24fps right?, lets build a 24hz monitor to gimp ourselves a bit more
![rolleyes.gif]()
Warning, this post might contain sarcasm, and other nasty bits, please consult an adult before readingOriginally Posted by 2shellbonus
Just to point out something.
Input lag on a 125hz mouse is 8ms. Then there is a certain delay between the switch being pressed and the that signal being converted to a data signal. That is another 8ms or so (usually less). Total lag is 16 ms.
Now if you take the best mouse on the market you will get 1ms input lag and 1ms signal conversion. Total delay 2 ms.
Great 8 times less. Then we add a human. Which is supposed to use the mouse. And human input lag (or reaction time) is 220-250ms.
So whatever the mouse you have the biggest problem and the biggest lag is you.
all the more reason to avoid even more "lag", or are you saying youd play the same with a 60hz monitor with 50ms+ gtg time, 60ms input lag, 30 fps ingame and a mouse with another 50ms or so input lag, why not play with 500ms ping while youre at it, i mean, you the human is the -only- limiting factor, right?
Wait, why even 60hz monitor, people can only see 24fps right?, lets build a 24hz monitor to gimp ourselves a bit more
Joined
·
3,097 Posts
Quote:
Only Keyboard PS/2 is interrupt based and really "better" in that regard than USB.
Mouse PS/2 is not interrupt based iirc. I think the standard polling rate is only 80 Hz even, can be overclocked to 200 Hz.Originally Posted by Offler
It also depends what you want to measure.
a) Mouse to PC
b) Mouse to Display
First choice is fully within responsibility of mouse manufacturer. Second one is affected by ... everything. Typical mouses are now USB only, but I still prefer PS/2 interface simply because mouse will receive its own interrupt (IRQ) and its not depending on CPU as USB mouses.
This is also why is measurement from blurbusters completely wrong if they wanted to measure Gsync parameters. The method measures difference between clicking the mouse and time when shooting appears on display. There is simply no reason to say that gsync monitor affects something when there is CPU, GPU, memory latency and absolutely everything in the way.
But yes, its true that you can use that method to measure two different mouses, but the results will me very inaccurate.
Only Keyboard PS/2 is interrupt based and really "better" in that regard than USB.
Joined
·
2,567 Posts
Every PS/2 mouse receives IRQ 12 over Slave PIC. Keyboards receive IRQ 01.
http://en.wikipedia.org/wiki/Interrupt_request
At least this was true years ago on my older PC. For now I simply know which chip on my mainboard handles PS/2, and apparently PS2 mouse still uses IRQ 12 in Windows 7...
When it comes to PS/2 connector i dont have info regarding polling intervals, and information on Wikipedia regarding this interface are badly sourced...
Since I start to use Logitech keyboards and mouses (KB 250 deluxe and MX 518) I never had trouble like "beeping" while mouse and keyboard stopped to work when I simulated CPU overload. USB devices usually can be affected by CPU.
What I really wanted to try out is to compare is mouse for COM port (serial port). Latest serial ports I have seen were able to handle up to 1mbit, while device connected on this port can be small computer of its own (such as raspberry). Usually mouses for PS2 are rare, and mouses for Serial ports are impossible to find.
http://en.wikipedia.org/wiki/Interrupt_request
At least this was true years ago on my older PC. For now I simply know which chip on my mainboard handles PS/2, and apparently PS2 mouse still uses IRQ 12 in Windows 7...
When it comes to PS/2 connector i dont have info regarding polling intervals, and information on Wikipedia regarding this interface are badly sourced...
Since I start to use Logitech keyboards and mouses (KB 250 deluxe and MX 518) I never had trouble like "beeping" while mouse and keyboard stopped to work when I simulated CPU overload. USB devices usually can be affected by CPU.
What I really wanted to try out is to compare is mouse for COM port (serial port). Latest serial ports I have seen were able to handle up to 1mbit, while device connected on this port can be small computer of its own (such as raspberry). Usually mouses for PS2 are rare, and mouses for Serial ports are impossible to find.
Joined
·
1,213 Posts
Camera might do the work but be sure not to test with the blurbusters method.
1- We want to test sensor lag, not clicks.
2-Animation is slower at playing than the "shot" being processed. It's not immediate feedback on input registration.
1- We want to test sensor lag, not clicks.
2-Animation is slower at playing than the "shot" being processed. It's not immediate feedback on input registration.
Joined
·
696 Posts
Quote:
Bigger values matter. Smaller ones dont. 50ms g2g is a big value. And count for up to 20% of the total reaction time. Ofcourse this will harm gameplay.
A 5ms difference is 2%. And this will never have any impact on your performance.
What you said is utter nonsense. And you completely missed the point.Originally Posted by Axaion
Warning, this post might contain sarcasm, and other nasty bits, please consult an adult before reading
all the more reason to avoid even more "lag", or are you saying youd play the same with a 60hz monitor with 50ms+ gtg time, 60ms input lag, 30 fps ingame and a mouse with another 50ms or so input lag, why not play with 500ms ping while youre at it, i mean, you the human is the -only- limiting factor, right?
Wait, why even 60hz monitor, people can only see 24fps right?, lets build a 24hz monitor to gimp ourselves a bit more![rolleyes.gif]()
Bigger values matter. Smaller ones dont. 50ms g2g is a big value. And count for up to 20% of the total reaction time. Ofcourse this will harm gameplay.
A 5ms difference is 2%. And this will never have any impact on your performance.
Joined
·
1,213 Posts
@Axaion
LOL 50ms gtg would require refresh rates broadly under 20Hz ^^
@2shellbonus
Mouse movement is a continuum, not a one time action hence the "feel", the overall perception of movement can change.
Will it give you lower scores in game? Probably not. Can it make playtime more enjoyable? Possibly yes.
Btw I'm about 25% less laggy than your claims
![wink.gif]()
LOL 50ms gtg would require refresh rates broadly under 20Hz ^^
@2shellbonus
Mouse movement is a continuum, not a one time action hence the "feel", the overall perception of movement can change.
Will it give you lower scores in game? Probably not. Can it make playtime more enjoyable? Possibly yes.
Btw I'm about 25% less laggy than your claims
Joined
·
2,269 Posts
Quote:
Quote:Originally Posted by 2shellbonus
What you said is utter nonsense. And you completely missed the point.
Bigger values matter. Smaller ones dont. 50ms g2g is a big value. And count for up to 20% of the total reaction time. Ofcourse this will harm gameplay.
A 5ms difference is 2%. And this will never have any impact on your performance.
My point is it adds up, and in a multiplayer eviroment if everyone else is equal skill and has 100ms lower delay, they will have a 50% reaction time advantage, which is huge.Originally Posted by the1freeMan
@Axaion
LOL 50ms gtg would require refresh rates broadly under 20Hz ^^
@2shellbonus
Mouse movement is a continuum, not a one time action hence the "feel", the overall perception of movement can change.
Will it give you lower scores in game? Probably not. Can it make playtime more enjoyable? Possibly yes.
Btw I'm about 25% less laggy that your claims![wink.gif]()
Joined
·
3,097 Posts
Quote:
What program is that in your screenshot?Originally Posted by the1freeMan
@Axaion
LOL 50ms gtg would require refresh rates broadly under 20Hz ^^
@2shellbonus
Mouse movement is a continuum, not a one time action hence the "feel", the overall perception of movement can change.
Will it give you lower scores in game? Probably not. Can it make playtime more enjoyable? Possibly yes.
Btw I'm about 25% less laggy that your claims![wink.gif]()
Joined
·
1,213 Posts
Quote:
http://www.draebenstedt.de/reaction/reaction.html
funny thing is the program actually has a function to measure inputlag based on putting your mouse against the screen with a moving image.
Too bad it doesn't get nowhere near consistent results, even taping CDs to the bottom of the mouse to maintain constant distance.
I don't know maybe with super high LOD it works ^^
ReactionOriginally Posted by Ino.
What program is that in your screenshot?
http://www.draebenstedt.de/reaction/reaction.html
funny thing is the program actually has a function to measure inputlag based on putting your mouse against the screen with a moving image.
Too bad it doesn't get nowhere near consistent results, even taping CDs to the bottom of the mouse to maintain constant distance.
I don't know maybe with super high LOD it works ^^
1 - 20 of 58 Posts
Join the discussion
Join now to ask and comment!