This is a mouse fix that professional gamers uses for accurate mouse sensitivity. One of the Cyberathlete Professional League's (CPL) members have made this simple registry change that will disable any mouse acceleration made by Windows XP.
Additional info: http://forums.urbant....html#msg155672 (same page)
Windows and Linux by default has the USB ports working at a polling rate of 125Hz (of 1000Hz that USB is capable of) resulting into a latency of 8ms response times. It is possible to change the frequency to 250Hz(4ms latency), 500Hz(2ms latency) and 1000Hz(1ms latency) to get better response times by reducing the USB polling latency.
For Windows XP users:
Download this zip file with 2 utility: download
- usbmrs11.exe -
This utility allow you to change the USB polling rate to run at a lower polling latency. For Windows XP 32-bit (Requires Administrator privileges).
- dx_mouse_timer_dialog.exe -
This utility allow you to test the USB polling rate. For Windows XP.
Use dx_mouse_timer_dialog.exe to test or your system is already using 1000hz USB polling for the mouse. Some mouse drivers already do this for you. Some drivers such as those of Razer already come with 1000hz polling (their marketing guys call this Ultrapolling).
To change the USB polling:
1) Boot to safe mode and use the usbmrs11.exe utility to set the polling rate to 1000Hz, reboot.
2) After reboot, run dx_mouse_timer_dialog.exe again to see at what rate your mouse is working.
3) Reboot to safe mode again, and set the polling rate to whatever your refresh rate is.. ie. if you got 479Hz, set it to 500Hz, if you got 849Hz, set it to 1000Hz. However, you could leave it at 1000Hz, but the higher the polling rate, the higher the CPU usage will be. This particularly can have negative influence for UrbanTerror as the game is CPU intensive.
[glow=red,2,300]Note: This MUST be done in safe mode, if not you will get negative results![/glow]
You should see a noticeable difference in mouse accuracy/response to movements; although it might take some time to get used to it as well, for the reason that most people are used to this 8ms "lag".
Take note that this doesn't work with all mouses. Some simply just won't perform better with lower USB polling latency.
An excellent read on mouse performance review can be found on esreality.com: link. Also a good reference to see how good your mouse can perform.
For Windows Vista users
Follow this guide on servegame.com: link
For Linux users:
Cute penguin, hot or not?
Follow this guide on linux-gamers.net: link
How to test for positive/negative acceleration on Windows
For this test it is required to use a game that uses DirectInput. This is _very_ important as a non DirectInput game is directly affected by the windows desktop and the mouse DPI. This means that if you run a screen resolution smaller than the DPI of the mouse, you will get results of FALSE negative acceleration. This happens because of the way the windows mouse input works: when you hit the boundary of the screen (which you will if you move very fast using 1800dpi on a 1024x768 or even smaller resolution) the mouse will reset to the center of the Windows desktop. Which gives a false result of negative acceleration.
UrbanTerror can be configured to use DirectInput, for both Q3A client and ioUrbanTerror client users:
Use "+set in_mouse 1" in the command line or "set in_mouse 1; in_restart" in the console.
1) Start UrbanTerror
2) Set a low sensitivity (lower than your normal sensitivity).
3) Aim your crosshair at a specific point like a corner or something that's easily recognizable.
4) Slowly move your mouse all the way across your mouse pad to the edge.
5) Move the mouse *fast* (as fast as you can) back to the starting point.
Now one of three situations have occurred:
1) The crosshair has overshot the point you picked out. This is positive acceleration
2) The crosshair has undershot the point you picked out. This is negative acceleration.
3) The crosshair lands spot on (with human error in mind). No acceleration is present.
As easy as this, and repeat the test 3-5 times to make sure it's a consistent result.
Oh and the final thing, if the crosshair suddenly stops moving like in a choppy movement and SERIOUSLY undershoots the spot, were talking about skipping (this should happen though, as speed tests show max speeds of 2.5 m/s (450 DPI) - 4.5 m/s (1800 DPI) which you will hardly ever reach using normal movements).
I hope this helps out.