1/7/2016 Update significantly shifts mouse sensitivity value

[kvbev]

OS: Arch Linux Desktop: i3wm Graphics: amd 7870 running radeonsi open source drivers and mesa3d 11.1.0

Mouse: Corsair M65 hardware set DPI: 800

Mouse acceleration has been disabled globally for my desktop using the libinput method described here: https://wiki.archlinux.org/index.php/Mouse_acceleration#Disabling_mouse_acceleration

Ingame raw input: off Ingame accel: off

before the update I played with an in game sensitivity value of 1.4

after testing in deathmatch with the new patch the sensitivity value I have to use to even get close to the original aim movement speed is somewhere between 2.4 and 2.8.

I could probably get a more exact value however it seems minor movements of the mouse are lost on input or feel more sluggish and even with higher sensitivity values, finer aiming control does not feel as it was before.

for me I would consider the game playable once I find a new value to use however the massive shift of how the sensitivity value is interpreted is unfortunate.

for me turning raw input on or off does not change anything.

[kvbev]

I would like to reiterate for people who have still not been able to find an appropriate new sensitivity value after both of these patches, and doubling your old value is too high, to instead increase your old value by half.

after playing yesterday and today where my old value was 1.4 playing after these patch's with 2.1 as my sensitivity is now exactly (within a mental margin of error) as it was before, also I may be more accurate now in aiming it feels like sometimes.

I do not have an explanation as to why this is, should it be accurate, but it may help others zero in on what their new sensitivity value should be.

this issue is now solved for me.

[hitcher404]

@kvbev what resolution do you play at? Can you try the old libSDL (a working binary is linked above by @Toqozz) and actually measure? Like physically move your mouse a specific distance and note the movement in game (pick a place on a map where you recognise the features maybe). I have done this with the old and new library and for me the sensitivity value exactly doubled produces the same distance moved in game.

@andras-kiss the error you get using your libsdl: SDL_Init(SDL_INIT_VIDEO) failed: No available video device is because you need the 32bit libX11 headers (and possibly others) installed when you compile it. I used the list here appending .i686 to each of them.

After testing with the old library (2-2.0.3) and new (2-2.0.4), as well as the patch @coffee-lord linked it seems the issue from #44 was indeed always present, but the new version of sdl made it more apparent.

EDIT: Please see the detailed update by ahnook to #44 with clear steps to reproduce and video demonstration.

[andras-kiss]

@hitcher404 Do you mean appending :i386? When I append .i686 it says

E: Unable to locate package mercurial.i686

When I try it with :i386, I get

The following packages have unmet dependencies: mercurial:i386 : Depends: python:i386 (>= 2.7) but it is not going to be installed Depends: python:i386 (< 2.8) but it is not going to be installed Depends: mercurial-common:i386 (= 3.1.2-2+deb8u1) but it is not installable E: Unable to correct problems, you have held broken packages.

With the latest update (1/8/2016) the bug still persists. I agree with @hitcher404 , the bug was always present, but was less noticable previously. I discovered it a while ago, when I just moved my mouse back and forth very quickly between the same positions on the mousepad, but the cursor kept creeping towards the bottom right direction. The bug is worse with lower refresh rate, resolution doesn't seem to affect it for me. Tested it with a Razer Deathadder 3.5G on 450dpi, 900dpi, 1800dpi, and 3500dpi on 125Hz and 1000Hz, a Microsoft Wheel Mouse Optical 1.1a 450dpi at 500Hz, a Steelseries Kinzu v2 and v3 at two different dpi settings (which I can't remember, but it didn't make any difference regarding the bug) at 500Hz.

I have a similar concern as @AtomiCharge about VAC. I think it might cause a VAC ban, and I really don't want to risk it with a 100+ game steam library.

[hitcher404]

@andras-kiss try just installing the 32bit versions of the lib* dependencies listed on the SDL faq. You dont need 32bit versions of the tools in that list like mercurial, make etc.

[andras-kiss]

@hitcher404 Success! The cornerbug is now gone, and I used @ahnook 's version from #44. But I think I will use the official valve version, because I don't want to risk a VAC ban. Let's hope Valve will patch their version. Can we request a patch based on #44? I think they read this thread don't they?

[jamezmoran]

IIRC, and don't quote me on this, VAC won't ban you if one of your libraries don't appear to be what is expected, it'll just throw you out of the server and say your files are invalid, even on Windows.

That aside, I tested the game today and I can't replicate the issues seen in this thread at any resolution. This appears to be fixed, thanks!

[ghost]

Could someone please explain what's the point of now Linux players having 50-100% more in the sensitivity setting than Windows players?

[jamezmoran]

From what anyone can tell, that is not the case at all. The previous behaviour is what was broken and meant people moving over from windows had to half their sensitivity values to get comparable mouse movement. This patch simply removes that buggy behaviour (libSDL2 firing double the amount of mouse motion events that it should). Whether you agree or not with having the sensitivity altered is a good idea, the fact remains that the previous behaviour was entirely caused due to CSGO being based on a buggy implementation of SDL2 and would most likely have been patched at some point in the future anyway.

[ghost]

@jamezmoran: Then how do you explain the fact that my sensitivity now seems to be around 3.3 with 400 dpi when it was a normal 2.2 before? What pro player plays with 3.3 sensitivity? It used to be the same as in windows, not anymore.

[craysiii]

If you feel it was the same in windows then I'm sorry to say you're probably mistaken. They have been quite different for some time and when I used to dual-boot I would switch between them even in the same day. I eventually started sticking to one version. The new patch has fixed all the issues for me and now it is similar to that of windows.

[FrancisBaileyH]

Has anyone figured out an exact formula to match your previous mouse sensitivity? I was playing at 2.14 sens and doubling it feels too fast, I'm really struggling here. After 500 hours on the same sensitivity readjusting is not working

[JoshuaMurphynz]

You can use the sensitivity command to easily change the sense.

Just a fyi as it easier to test this On 11 Jan 2016 22:28, "Francis Bailey" notifications@github.com wrote:

Has anyone figured out an exact formula to match your previous mouse sensitivity? I was playing at 2.14 sens and doubling it feels too fast, I'm really struggling here. After 500 hours on the same sensitivity readjusting is not working

— Reply to this email directly or view it on GitHub https://github.com/ValveSoftware/Counter-Strike-Global-Offensive/issues/574#issuecomment-170483475 .

[andras-kiss]

@FrancisBaileyH I've found that raw_input is broken, so you have to manually disable mouse acceleration. I use

xset m 0 500

in my .xinitrc. Try to play with the settings in-game, switching them on and off. Have you played 2.14 on linux or windows? Previously you had to double your windows sensitivity in linux to get the same sensitivity. But with the latest patches, it is the same across all platforms. As I understand, the problem was SDL2, every mouse movement event triggered two coordinate updates, now it is fixed. I get the same mouse handling as previously by doubling my pre-patch sensitivity.

[jamezmoran]

For me it is almost exactly 50% higher. The mouse motion still did not feel exactly the same as it did before, actually it felt a bit finer, but I've been on DM and Faceit this weekend and I'm well adjusted now.

[FrancisBaileyH]

I exclusively play on Linux as I was never able to get my sensitivity matched up with linux. @adras-kiss thanks, I'll try adding that to my .xintrc

[FrancisBaileyH]

Perhaps I'll just double it and practice in dm as much as possible

[FrancisBaileyH]

Or 50% would mean adding half of the current value not double

[jamezmoran]

Note that putting that in your.xinitrc will only work if the desktop you are using even executes it, I don't believe GNOME 3 does.

[jamezmoran]

When I said 50%, I meant 50%. My sensitivity was 0.7 before, now it is 1.05. Different users are reporting different values that are optimal.

[hitcher404]

I'm not sure what you guys are experiencing, but outside of the resolution based differences discussed above, exactly doubling your sensitivity means it is identical to what it was previously. Again as mentioned above, please feel free to test with the old version of libSDL. Also, what it 'feels like' is a poor metric, try measuring the distance your mouse moves vs angle turned in game.

[ghost]

"Feels like" is not really a poor metric. I have over 800 hours in CSGO and I can determine the difference to be significant without any doubt. It clearly shifted a lot with the latest update.

[hitcher404]

When there is no acceleration applied the distance your mouse moves directly translates into in game movement. If you cover the same angle, while moving your mouse the same physical distance, your sensitivity is the same.

[salixh5]

Anyone having a problem with the mouse cursor sometimes not going to the left correctly? Like, it doesn't seem to aim completely smooth all the time. I don't seem to have this problem with my mouse outside of CSGO at all.

[JoshuaMurphynz]

Yes this has been reported here on github On 13 Jan 2016 03:05, "Jei Four" notifications@github.com wrote:

Anyone having a problem with the mouse cursor sometimes not going to the left correctly? Like, it doesn't seem to aim completely smooth all the time. I don't seem to have this problem with my mouse outside of CSGO at all.

— Reply to this email directly or view it on GitHub https://github.com/ValveSoftware/Counter-Strike-Global-Offensive/issues/574#issuecomment-170921393 .

[ghost]

This was fixed in January (please close).

[Tele42]

Indeed, this was the expected result of https://hg.libsdl.org/SDL/rev/957a6ed39ad0 in libsdl2 getting into a production release.