Getting mouse and keyoard input

Hey GameDev, at the moment in my game/s I've been using GetAsyncKeyState(xx) to get keyboard input and GetCursorPos to get mouse data. What's the difference between this and say using RAW input or DirectInput? Should I bother changing or is GetAsyncKeyState/GetCursorPos fine? I'd rather not spend too much time on getting input, hency why I use these rather than RAW input

If they work fine for you, I don't see any reason to change, regardless of what features other solutions offer.
I wouldn't bother with directInput/xinput either way, unless you want joypad support it won't give you much.

If they work fine for you, I don't see any reason to change

I like that. But I was just wondering what one offers compared to the others. In regards of speed, smoothness etc

You can't beat your current method for simplicity. If it works then stick with it. If you ever decide it's no longer enough, Raw Input is pretty simple to get up and running; easier than DirectInput.

Using and handling the windows messages is probably the fastest option as the raw messages are rapid and you get them for free anyway. When I say free, I mean they are sent regardless of other solutions, so you might as well use them. If dealt with in a nice way, then you get very smooth results.

Basically the way I do it is pretty simple:

1) Set up a keyboard class or manager
2) Set up an array of 256 keys, giving them a state of "not pressed"
3) Once a WM_KEYDOWN message is sent, I convert the wParam into my set of key defines "static_cast(wParam)" which gets sent into a helper function within the keyboard class to register that key as "pressed", if another message is sent for the same key and that key has a "pressed" state, then that key is set to "held"
4) Then once the WM_KEYUP message is sent for that key, set it's state to "not pressed"

Then in the main game code, it's just a case of simple checks like so:

if ( keyboard.KeyHit( Key_Escape ) ) // kill the program

if ( keyboard.KeyHeld( Key_Space ) ) // accelerate

Mouse clicks can also be used exactly the same using the WM_LBUTTONDOWN, also get access to the mouse's location.

So yeah, my opinion - Raw data is very easy to set up, and is very fast. Once done correctly, it can be very simple, easy to use/read.
Using the raw data, you can also design your own interface, instead of using/learning another like Direct Input.

Personally I wouldn't use Direct Input, it can be slow and implements an unneeded layer between the messages. Also it needs to be set up, and can be messy. For example, you have to check the keyboard hasn't been lost every frame, such as clicking on a different window, means the input needs to be acquired which is annoying.

Yeah [color=#282828][font=helvetica, arial, verdana, tahoma, sans-serif]

[background=rgb(250, 251, 252)]GetAsyncKeyState/GetCursorPos [/background]

[/font]is very simple to use, but it's limited.

WM_KEYDOWN/WM_KEYUP and WM_LBUTTONDOWN aren't raw input messages, those are the standard input messages. They give you the same exact data that you get from GetAsyncKeyState and GetCursorPos.

Raw input gives you unfiltered input from the mouse and keyboard, which means Windows doesn't do anything to process it before giving it to you. So for instance with raw input you don't get a cursor position, just get data telling you how much the user moved the mouse. If you wanted to implement a cursor with it you'd have to do that yourself, and you'd probably end up with a pretty wonky cursor that doesn't match the Windows cursor position. However if you were implementing something like an FPS camera, then raw input is nice since you get to use the full sensitivity of a high-DPI mouse. You also can implement your own custom filtering that's suitable for your camera setup.

Oops, I miss understood "raw data", thanks for clearing that up!

Raw input is for high-definition devices. I usually ignore implementing raw input devices and just stick with the regular keyboard and mouse checks using windows message proc.

If you are just hard coding a simple game, then do what you will. But my opinion is that if you are creating an engine that your game will be using, you better do it right the first time around using Raw Input otherwise you will be kicking your self in the butt at a later date.