CS:GO - m_rawinput vs rinput

[nicolovbg]

Hello!

 

I decided to try your program in Quake Live. I play with in_mouse -1 (=m_rawinput 0) instead of in_mouse 2 (m_rawinput 1) because I've discovered (of course by feeling; but I am such a sensitive person to this stuff) that RawInput in QuakeLive has relative delay as well.

 

So, what's the problem with your program? Well, I start the game, play a little bit, open Rinput v1.37, type quakelive_steam.exe, the *.dll file successfully injectes, I go back to the game, play a little bit (everything's fine), then press ESC. After that buttons in the main menu cannot be pressed, like MOUSE1 doesn't work at all. Then I press ESC again and go back to the game.

 

And here the real problem startes. Eventually, I can shoot, but I cannot aim with my mouse, the crosshair doesn't move. Could you please explain the problem to me, or, try to fix it?

 

Thank you

[DPI Wizard]

Rinput is not my program, I have nothing to do with it. I just tested it

But in CS:GO, the mouse does not work if you are using Rinput and have raw mouse input turned on, maybe it's the same problem with Quake Live. If it's already off, then I don't know.

[Rud]

 

Video

This video visualizes the discrepancy:

 

 

Can you make the same video CSS?

[hellothere]

Thanks for this! Have you tested MARC mouse fix though? Should I use that?

[DPI Wizard]

Thanks for this! Have you tested MARC mouse fix though? Should I use that?

If you use raw mouse input, it won't affected by this fix. Raw mouse input is pretty much perfect if you want real 1:1 mouse-to-game movement, no need to mess around with anything in between. 

[Xbye]

Just thought I'd throw this in here. I got a new mouse (Zowie FK2) and the "jitter" i seemed to deal with from the G502 and the G402 while using SourceGL in Left4Dead 2 seems to be gone(or very minimal).

 

Furthermore, I'm curious as to what different forms of mouse calculations are being used by different engines. For example; I use 4/11 windows and play with rawinput or sourcegl in L4D2 (source-engine) as using rawinput 0 in L4D2 results in a very, very low amount of movement (due to my 4/11?). However I can turn off rawinput and play without sourcegl/rawinput in NS2(spark engine) just fine.

 

What calculations are going on in a game like NS2 that allows me to turn off rawinput but still have the same mouse configuration settings as having no rawinput?

[XCNCTX]

Hello. I was wondering if anyone here could tell me about the pack loss of the new rinput 1.44? I have searched everywhere to see if someone has tested it. Are you planning on testing the packet loss of rinput 1.44? Thanks

[Drimzi]

Edited January 28, 2018 by Drimzi

[Vols and Jezuz]

The last time I tested was Dec 9, 2015. My results were 0/500,000 input packet discrepancies at each of fps_max 128, 425, and 999 with RInput v1.41. The newer releases were more about recoding and fixing some bugs in games' menus, so v1.44 should still be equivalent to m_rawinput 1.

That is indeed the newest release of sourceGL. It was taken down from Sourceforge because it was not open source, right as I was just starting a full open source recode in C++. But alas, I've since gone back to school, and simply don't have time to work on this project for the next 2 and a half years. Maybe at that time I will be able to do the full recode.

Hope this helped.

[DPI Wizard]

I'll do a new test some time with physical mouse movement rather than emulation now that I have a machine that can do this.

[hogaNmTw]

Hi!  : ] 

Thanks for all your Hard work and time!

I would like to know your thoughts and understanding please on the latest version of RInput 1.44 found here:

https://fearless-assassins.com/files/file/1656-rinput/

[wazer]

On 10/26/2017 at 4:41 PM, DPI Wizard said:

I'll do a new test some time with physical mouse movement rather than emulation now that I have a machine that can do this.

Is it time for an update and comparison with forked 1.44 version?

 

https://www.reddit.com/r/GlobalOffensive/comments/i5ddor/csgo_mouse_input_flawed_whitelist_rinput_144_or/

[DPI Wizard]

1 hour ago, wazer said:

Is it time for an update and comparison with forked 1.44 version?

 

https://www.reddit.com/r/GlobalOffensive/comments/i5ddor/csgo_mouse_input_flawed_whitelist_rinput_144_or/

Here's the issue with anything claiming to be better than rawinput... It's already "perfect". You can't possibly improve on anything that correctly processes the raw input because you can't possibly have any more information. 

Any addon/mod/program that claims to be "better" has to get the mouse input from somewhere... and guess what: it's the raw input. So as long as the report from the mouse, say "move 100 counts right", reaches the game and it does exactly this, there's literally nothing to improve. The only issue is when this report does not reach or register with the game, which in CSGO is pretty much a non issue. A lot of games do struggle with this though, but CSGO is not one of them.

Also, any additional software between the mouse and the game will only add latency. While it might be insignificant it certainly outweighs the potential benefit of improving something that doesn't really have an issue.

Maybe there's some benefit to certain older hardware I'm not aware of, but if everything is up to speed this isn't something to even consider.

I don't think I can test the latency specifically other than by feel for now, but accuracy I can definitely test. So I will do this and see if there's any benefit in a real world environment. It would surprise me if it is, but I'll give it a go and eat my words if I'm wrong