Things you should know about Windows Input, but would rather not
ph3at.github.io
ph3at.github.io
I have secret desires to convert one of my old MacBook Pro devices into a Steam-via-Proton setup to get out from under the tyranny of Windows but since gaming is designed to be a break from "work" ... that's why it's not already done
Unless they've fixed this in the last few years, switching your input mode like that while World of Warcraft is running will cause some kind of crash that prevents you from using the in-game chat. This felt especially egregious because switching input modes is the kind of thing you want to do all the time if you're using the in-game chat.
Option 2: launch game, navigate to settings, write down on a sheet of paper what the current keybindings are (since one I press "." for "e" it's going to either whine or blank out whatever "." was mapped to when I started), repeat that exercise about 25 times per game, times the 15 games I have in rotation right now, feel that was a great use of my "downtime"
Option 2 has the added bonus of making it 50/50 whether the game help text knows I remapped, and thus says "press . to open door" or whether it continues to say "press e to open door" and I have to guess whether it means "their e" or "my e"
There's some convention for which scan code to use for which physical position on a keyboard, but that's not correlated with what's actually printed on the key caps. E.g. on common QWERTY keyboards the "A" key will have scan code 4, but on AZERTY keyboards it will have scan code 20.
Games can probably get away by listening for the scan codes of the keys in the position commonly used by WASD, but it's a bit fragile, and they can't actually tell you what's printed on the keys they're listening to. The lack of consistenency is certainly annoying, though...
Games SHOULD use scancodes for positional mappings (like WASD) and rely on the system keymap to decide what letter to display them as. There is no "probably" here, the scancode is the scancode and the keymap is the keymap.
Games OFTEN use the character codes directly regardless if it's for a positional mapping or not. This requires explicit support in the game for each possible system keymap, otherwise you end up with nonsense mappings.
The operating system knows how to map scan codes to characters based on the keyboard mapping the user has selected.
Win32: MapVirtualKeyExA / MapVirtualKeyW
MacOS: CGEventKeyboardGetUnicodeString / UCKeyTranslate
Linux: xkb_state_key_get_utf8
Stop reading character codes directly, use the scan code and when displaying them map it using the system API so people see the right freaking binding. It's not rocket science,
However, we also reflect your current key bindings with dynamic button prompts that tell you which key to press in game. For this we translate the determined physical key back through the selected keyboard layout to figure out what the corresponding label on your keyboard might be.
Most of this is just win32 API shenanigans, but the following post provides a bit more detail on reading inputs this way.
https://blog.molecular-matters.com/2011/09/05/properly-handl...
The fun parts you might not know is eg. that if you want to ship on steam you will need legacy messages for its overlay to function. Because its a web based thing ^_^
Practical solutions usually include either letting user to pick their perfect input type and assume they know limitations like no windowed mode. Or expecting users to reduce 8k to reasonable number in mouse software. Users already know how to read forums or reddit and games that do now work fine with 4k/8k polling rates are well known and numerous.
There is an alternative that my colleagues going to explore = make it Microsoft problem. There is Microsoft provided library that can do input for you ...
After working on one a bit I would say library way is not good enough :) but it is sure thing cost effective way to support a platform. Or even few platforms shared by same vendor.
It’s very Linux specific but can frame by frame manipulate the input that an application is getting, meant for speed runs of games.
GameInput does everything these older APIs do without having to mix them all together or use a bunch of them to get the information you want
You still need legacy input for Window interactions, and the message queue still gets flooded by fast-polling mice.
> When the hardware mouse reports an interrupt, indicating that the physical mouse has moved, Windows determines which thread should receive the mouse move message and sets a flag on that thread’s input queue that says, “The mouse moved, in case anybody cares.” (Other stuff happens, too, which we will ignore here for now. In particular, if a mouse button event arrives, a lot of bookkeeping happens to preserve the virtual input state.)
> When that thread calls a message retrieval function like GetMessage, and the “The mouse moved” flag is set, Windows inspects the mouse position and does the work that is commonly considered to be part of mouse movement: Determining the window that should receive the message, changing the cursor, and determining what type of message to generate (usually WM_MOUSEMOVE or perhaps WM_NCMOUSEMOVE).
The book version of Raymond Chen’s blog[2] elaborates on the subject of messages arriving in an empty forest:
> One of [the] special messages is the WM_MOUSEMOVE message. As you saw earlier, the message is not added as an input message when the mouse moves; instead, a “the mouse moved” flag is set. When the window manager goes looking for an input message and the “the mouse moved” flag is set, it clears the flag and generates a WM_MOUSEMOVE message on the fly, adding it to the list of input messages (or coalescing it with an existing WM_MOUSEMOVE message).
> Other special messages that fall into this “generated on the fly” category are the WM_PAINT, WM_TIMER and WM_QUIT messages. The first two messages are generated even later in the message search process, only after no applicable input message list was found, and only if the message filter indicates that that type of message is being requested. (The WM_QUIT message is even shier than the paint and timer messages, because it emerges only after the posted message list is empty. On the other hand, the WM_QUIT message is also bolder, in that it ignores the message filter.)
Keeping that in mind, using PeekMessage() and thus materializing the pending WM_MOUSEMOVE should be done very carefully (though with PM_REMOVE it shouldn’t be a problem). In fact, it seems like unless you care about WM_TIMER, WM_PAINT, or WM_QUIT, the right thing to do upon seeing a WM_MOUSEMOVE (but not a WM_NCMOUSEMOVE) would be to behave as though the message queue is currently empty and proceed to do whatever other processing you need to perform for the current frame.
Still, I don’t get why this should be a problem. If you receive a WM_MOUSEMOVE and your only action is to discard it, surely it doesn’t take 125 μs to do that? And if it doesn’t, you shouldn’t see more WM_MOUSEMOVEs that can only get generated at 8 kHz at most.
Suggested experiment because I don’t have a fancy mouse: if you insert
if (msg.hwnd == <root window> && msg.message == WM_MOUSEMOVE) continue;
[1]: https://devblogs.microsoft.com/oldnewthing/20031001-00/?p=42...
Many, but not all recent game controllers come with accurate and stable enough calibration so that you don't need to do this very often or at all. It's mostly older hardware that needs to go through the process. Most 10 year old controllers (sticks/wheels etc.) can still be perfectly usable today after calibration!
"Find a setting" > type "joy" suggests "Setup USB game controllers". Lists the controllers/joysticks. Click "Properties". Click "Calibrate..."
Reading keyboard arrows or WASD has timing problems. Although the program gets key code up/down events, the key repeat software seems to interfere with the timing of the events and slows them down. Software repeats can be ignored, but still cause problems. Holding down a repeating key for long periods can cause lost events. X11 is especially bad, but there seem to be some problems with Windows, too.
Another thing nobody has fixed in the Rust game dev ecosystem yet because the developer base is tiny. Rust needs one good first person shooter project with enough push behind it to wring out the performance problems in interactive 3D.
It's been on my mind to try the google IME for some time, to see if it fixes this issue, but haven't got around trying it yet.
[1]: https://learn.microsoft.com/en-us/windows/win32/api/winuser/...
[2]: https://github.com/jchv/WmPointerDemo
> Disable legacy input, create your own legacy input events at low frequency. Another one in the category of “silly ideas that probably could work”, but there are a lot of legacy messages and this would be another support nightmare.
This was my first thought when reading the article. Frankly it would be a pain, but this seems like the sort of thing that someone only has to do one time.
(Calling them "legacy" messages seems weird -- this is the normal current way that input messages work for regular applications).
Will this matter? I'm not quite sure.
I think Microsoft's WebView2 component enables Mouse-in-Pointer mode using the undocumented syscall for doing it per-HWND, but apparently Raw Input is per-entire-process, so if you aren't generating "legacy" events, will a WebView2 component in your process function as expected? Maybe if you go the synthetic pointer route, but again, that's a pretty big can of worms in and of itself.
Windows input is indeed very tricky. Seems like there's only one way to figure out some of these answers. Perhaps I should add RawInput to my test application and find out what the interaction is.
[1]: https://learn.microsoft.com/en-us/windows/win32/api/winuser/...
I don't think this is silly at all. People are too timid about using multiple processes in general. It's OK, there's nothing wrong with it! We ought to have more libraries to help make multiprocess apps.
Is this why a game renders only a few frames per second and slows down in time when my wireless Xbox controller disconnects due to low battery? I always assumed it was some mechanism to make the gameplay “fair” if your controller had an intermittent disconnect.
1) You should be processing the whole message queue for all types of messages in one go. When using raw input, author is peeking for "not input" event types. To me, this seems suspicious for performance. With a raw input device registered, the message queue is going to be huge.
While I don't know the underlying data structure of the message queue, if we assume an array, then the author's code will do a while loop of O(N) lookups for "not input" messages and remove them.
The correct approach is to dispatch the whole queue and ignore messages of types that you don't care about in your message handler.
2) You SHOULD be using the legacy input events for games, not raw input. The OS input massaging (ex: acceleration) is something that should remain consistent between applications. If there's acceleration: that's what users expect. You don't want the mouse to suddenly behave different in your application. Your games tuning values (sensitivity, accel) should apply relative to the OS values.
The author claims that this generates an overwhelming number of events for a high-performance mouse. Is this not the case?
In those type of games, you aren’t controlling a cursor, so there’s no “expected behavior”.
For more critical observers, you're right: raw input is preferred. However, most users (casuals) expect similar settings as their OS. It's probably best left up to the user to decide, but to default things to the OS legacy settings. A more advanced player will know to tune the settings while the less advanced just want to launch and go.
For everyone else, there's the OS values.
We're not having the same argument, but I didn't clarify in my initial post properly: It's best to launch your app for the first time with the default OS settings. Let the "power users (fps gamers)" tune settings that enable raw input.
[0] https://research.nvidia.com/publication/2022-08_mouse-sensit...
I loved the Middle-earth: Shadow of Mordor game but when I tried to play its sequel, Middle-earth: Shadow of War, I found the mouse input absolutely horrible.
Horrendous mouse lag, floaty camera control, and all the signs of a "mouse emulating gamepad" type situation.
Does anyone know if there's a way to fix this game? I'd love to play it on PC with a mouse but can't stomach it as-is.
btw, I never played this particular game but lots of others have this issue.
I would say think of it this way: what happens when the mouse moves between USB hid messages?
If the mouse moves between messages sent via USB HID do the later messages capture the delta between the last and current message. Including the movement that happened in between?
Do the input messages, both those sent over USB by the mouse as well processed and presented by Windows, give relative or absolute positions?
The messages already represent a discrete sampling if a signal. You should be able to recreate the original information reasonably well without processing every single message.
> This would at least solve the problem for the vast majority of users, but it’s not possible, as far as I can tell. You seemingly cannot switch between legacy and raw input once you’ve enabled it. You might think RIDEV_REMOVE would help, but that completely removes all input that the device generates, including both legacy and raw input.
And maybe it will have some missed event between your ticker for processing game logic and fps.
I find it next to impossible to type "Now" really fast without having it come out as "NOw" much of the time. (Why I'm using this example is that it's the first word of the "Now is the time ..." test sentence).
The timing required is strict. The Shift key must be down before the letter key is down.
Keys activate strictly and immediately on the down event.
Note that virtual buttons on most contemporary GUIs, clicked by mouse, do not work this way. You must click and release the button for it to activate. If you click OK on some dialog, but change your mind while still holding down the mouse button, you can back out of it, by moving the cursor out of the button's area. You also cannot click down outside of the button, then move into the button and release. The button must receive both mouse down and mouse up, without the pointer leaving the area, in order to activate.
I'd like a mode in which the down events of non-modifier keys are deliberately delayed by some configurable number of milliseconds, so that when the user hits a modifier key at around the same time, and that key lands a little bit late, it is still taken into account.
It would also be nice to be able to experiment with a mode in which keystrokes are generated by their up events, not down, and the modifier is sampled at that time.
Your keyboard has a serial interface to your computer. Events are generated serially. If you press one first it will register first.
> keystrokes are generated by their up events
We don't do this because then you can't have auto repeat.
While that is obviously so, there are two events: a press and a release event.
The thing I run into is that the release time: Shift being held over too long, into the start of the next character. Now comes out as NOw because the up event for Shift comes after the down event for w.
So I don't think my proposals are necessarily the right ones.
Perhaps this: when a shifted letter is being typed (the down event came in a shifted state), then the rule should apply that an up event should be received to generate the keystroke, and that up event should still be in the shifted state.
Refinement: when a shifted letter is being typed in a situation where the Shift was held over from a previous shifted letter, then apply the logic. Thus N is not delayed; it comes out on the downstroke. But if Shift is still held while o is typed, that is a held-over Shift, and so the rule applies; the state machine must see an up event for the o and then decides whether it is shifted. If Shift is released before the o, then o is issued on the down event as usual.
Another idea: when a shifted letter is being typed in a Shift-holdover state, then issue the keystroke for the letter either on the release of the Shift, or the release of the letter, whichever comes first. If Shift is released first, assume it was a holdover and don't capitalize the keystroke. If the letter is released first, issue the capitalized keystroke. In the situation where the intent is to type Now, this will hardly result in a noticeable delay for the o, since the Shift hold-over period is very brief.
> you can't have auto-repeat
Auto-repeat is an arbitrary behavior that kicks in due to a timer going off while a key is continuously held down.
It generates multiple fake keystrokes even though there aren't multiple down events.
So of course you can still have it generate multiple keystrokes even though there aren't multiple up events (including generating the first keystroke in spite of there not being any up event).
Generation of keystroke based on the up event, beside been incompatible with repeating keys for long strokes, will slow down typing significantly, as it requires tracking timing pressing keys for longer duration. I'm pretty sure that this isn't only effect of me being used to track keypress timing on the way down, but an unavoidable result of the duration of the action.
Waiting for up event on contemporary GUI, when the contempt UI is a sluggish fit-to-nothing dirty touchscreen in a public kiosk is sensible. When you know an interface will yield more errors than intended input it is only sensible to assume that any input is a mistake unless the user is making an effort to validate it.
A game written for an IBM PC and everything after that can know exactly which keys are being held down and when they are released by intercepting the "scan codes" (or abstract keyboard events in a GUI event loop).
All that is missing is synthesizer-like velocity and pressure info. :)
> will slow down typing significantly
Only for people who have to look at the screen. :)
The hunt-and-peck beginners who look at their fingers are not effect, and neither are those who can look at something else or close their eyes.
A serious concern affecting even those people is that using the release event could reorder things, causing mistakes. Like say if someone types the sequece wh (involving two hands) such that they release the w later than h.
Why is this an issue? Your mouse is sending raw input 80,000 times per second to a screen that can only display 120 pictures per second.
Kinda sounds like an arbitrary problem that people who believe 8k mouses are necessary would worry about.
8 thousand, not 80 thousand. Many screens can display (much more) than 120 frames per second - My monitor goes to 240hz, there are 300hz displays available.
Rendering rate is very often not tied to game update rates - games can update their state at significantly higher rates than they display. A decent number of games have a "high level" skill of doing things like "pre-shooting" where the winner of a fight is done by figuring out "who realistically did it first".
Assuming that you can poll at 125hz is wrong.
> Why is this an issue? ... Kinda sounds like an arbitrary problem that people who believe 8k mouses are necessary would worry about.
People have these devices and run our games. We (unfortunately) need to handle them, just like a website needs to handle a poorly behaved client. Also depending on _when_ you sample, (i.e. which 5 messages you handle), you can get incorrect information from the device.
The core of the issue is trying to wrap polling style of logic into an event based system (which windows is at its heart). There is no real pretty way to fix it.
Well said. The irony being that the event based system on windows is wrapping the polling of the USB device in the first place.
It occurs to me that it may be possible to disable the wm messages. There is probably a callback hook location. Set that to 0 and it would probably not send anything. But that is just a guess and I would have to dig around in the docs to see if it is there. This indicates it is possible https://stackoverflow.com/questions/57008742/how-to-discard-... Then it would be up to you to use one of the other raw APIs to find out what is going on. https://learn.microsoft.com/en-us/windows/win32/winmsg/mouse... https://learn.microsoft.com/en-us/windows/win32/winmsg/about...
It’s not a high-end mouse.
Yeah, yeah OK polling rates, etc. Whatever. But honestly, PC gamers have been mostly fine for over 20 years of dealing with mouse input. Some major engines didn’t even support raw input until after the mid-2000s.
You can put a V8 on a Corolla… but if the curb weight changes and the torque eats up your tires, well, what did you expect? It’s just idiot engineering.
Neither of these things are "high-end" anymore though. You can get decent high-res mice for $50 and 120 Hz monitors for $250.
Now you may go back to bottom fragging.
This is a neat thing to see at the top of an article, very in line with the hacker spirit.
Sorry, I thought you meant 8k like the resolution, not 8kHz mouse sampling rate. I thought about deleting my comment, but it made me laugh so maybe it will others, too
I don't think I've seen any AAA games that allow me to run them in windowed mode with decent performance.
Valorant, for example, switches to windowed mode when I press Alt-Enter by mistake, but doesn't allow me to interact with the window much. It also locks up completely when Alt-Tabbing, so I don't know what's going on there.
Plus, I think the general expectation is to always run AAA games in full-screen mode. In such cases, disabling legacy messages is a viable approach.
You get full-screen performance with the ability to easily alt-tab without mode changes.
[1] https://learn.microsoft.com/en-us/windows/win32/direct3ddxgi...
Strangely enough, Windows 7 also supported the flip model, but only as an extension to D3D9, not 10/11.
I think the difference is between when you tell the game to run that way in it's settings, versus abruptly forcing it by doing alt-tab. Games where their default 'full screen' is really 'borderless windowed' don't have problems when you alt-tab, but a lot of them act weird when their 'full screen' actually is a true full screen mode and you force them out of it by alt-tabbing.
I remember playing CS:GO (and later, CS2) and it was super common to see cheaters in almost every other match.
In contrast, I haven't seen a single cheater yet in a match of Valorant.
You can see the difference yourself by checking the UnknownCheats forums for CS2 vs Valorant: the CS2 subforum is full of new hacks every day (most of which seem to be slight modifications of existing ones) while the Valorant one is completely dead.