stick mapping mouse failure

[ventus]

Yes, it sounds possible. Do you need more points for stick mapped to mouse?

What about 8 points? Will it be enough?

8 points is Ok, and I think you can consider a power function curve adjusted by a parameter, like titan fall 2. Use a slider to adjust the curvature of the curve, it will be much more smoother. And 8 points can exist at the same time if someone want a different curve. Let people choose which one to use. Thanks!

[reWASDer]

I believe that curves in Titanfall are almost the same thing that the presets you have in reWASD. Our presets are based on traditional behavior of Xbox controllers, while with Custom option, you can change it the way you want. With 8 dots, it will be much more precise

[ventus]

I believe that curves in Titanfall are almost the same thing that the presets you have in reWASD. Our presets are based on traditional behavior of Xbox controllers, while with Custom option, you can change it the way you want. With 8 dots, it will be much more precise

Yeah, waiting for your mouse simulate fixed and the 8 point! Good luck!

[ventus]

Yes, it sounds possible. Do you need more points for stick mapped to mouse?

What about 8 points? Will it be enough?

when will the next version be released? It has been august.

[Star-Lord]

Hello,

The new version expecting to be released in August. But, this release will not contain additional dots for the curve as the release is already full of new features and improvement.

Stay tuned!

[reWASDer]

Hello! We have released reWASD 5.5, and the issue with weird stick-to-mouse movements is solved. Please check the new version :satisfied: Hope to get any feedback from you.

The full changelog is here: https://forum.rewasd.com/forum/rewas...-and-many-more

[SlimeQ]

I believe that curves in Titanfall are almost the same thing that the presets you have in reWASD. Our presets are based on traditional behavior of Xbox controllers, while with Custom option, you can change it the way you want. With 8 dots, it will be much more precise

This isn't true at all. It's nearly impossible in ReWASD to get a smooth exponential curve because of the points system. The response curves in the Xbox Accessories app have the same issue. With 8 points you may be able to get closer but having to place the points manually is a really bad solution to this problem and will cause irregularities in the input.

In Titanfall and Apex Legends you get a "Response Curve" slider which essentially just sets an exponent. So your input magnitude is equal to deflection^n, where n is the slider value. I believe the default value is around 2, which gives you a nice smooth curve. A value of 1 will be linear. Personally I've finely tuned this value to give me precision control at low deflection while maintaining a higher sensitivity at the extremes.

I believe this is pretty standard in modern FPS games, although most games will not allow you to modify the exponent value. This really is the missing piece of the puzzle for kbm emulation in ReWASD to be a viable option for games.

[SlimeQ]

Just tried the new version with the manual curve inputs. I was able to get a decent solution though it required an open notepad window, a calculator, and 20-30 minutes.

These are the notes:

max = 32768

p1.y --> 0.2^2 --> 0.04 * max = 1311
p2.y --> 0.4^2 --> 0.16 * max = 5243
p3.y --> 0.6^2 --> 0.36 * max = 11796
p4.y --> 0.8^2 --> 0.64 * max = 20972

interval = 6553.6
p1.x --> 0.2 * max = 6554
p2.x --> 0.4 * max = 13107
p3.x --> 0.6 * max = 19661
p4.x --> 0.8 * max = 26214

As you can see, pretty complicated but I did manage to get a nice curve out of it.


I think I could get a slightly more accurate curve by spacing the X values exponentially instead of linearly but that will require some more tinkering. I may just write a script to do this calculation quickly.

One major issue I had inputting the values is that there are constraints placed on the values that interfere with typing the numbers in. The solution I came up with to get around this was to set ALL x and y values to 0 and then input them backwards, from point 4 to point 1.

[Star-Lord]

Hello,
Thank you for your feedback.

Glad that you're like it. As for manual inputting, N + 1 point on X or Y must be greater than N point, i.e all points should go increasingly. But the way you made is also elegant

Have a great gaming experience.

[ventus]

Hello,
Thank you for your feedback.

Glad that you're like it. As for manual inputting, N + 1 point on X or Y must be greater than N point, i.e all points should go increasingly. But the way you made is also elegant

Have a great gaming experience.

Hello, sorry to it's me again. My problem this time is that I always feel that when I push the stick a small distance, it is faster than pushing the same distance in the game using controller directly.So I simply did some tests. I used rewasd in Tarkov and compared it with Battlefield 1 and Apex. Both games use a linear curve, and tarkov with rewasd also use linear curve.But it is difficult for me to control the precise numerical experiments manually, so I only did rough experiments. This proves that when the maximum speed of the three games is adjusted to the same(When the stick is pushed to the maximum, the speed in the game is the same), the linear simulation of rewasd is indeed faster when the input is small. And when the dead zone is set to 8%（it is 2621 in rewasd）, rewasd gives me a completely different feeling from the other two games, and it feels much smaller. Based on these two points, I make the following guess. The game that directly uses the joystick is based on the distance pushed by the finger, while the joystick input used by rewasd is based on the angle, because when the joystick is pushed to a small degree, it may rotate 1 degree by pushing 1mm, but when it is fast When pushing to the end, you may need to push 3mm to rotate 1 degree, and there is a tangent conversion between angle and distance. This also explains why rewasd gets faster results when pushing the joystick slightly. For games that natively support controllers, the input signal has been converted, so push the stick slightly more slowly. But these are just my guesses. Can you use Battlefield 1 or Apex(linear curve) to do some accurate numerical tests for comparison based on this theory? If you can, can you try to adjust the input method to the same as these games? Thank you very much!

[reWASDer]

Hello, sorry to it's me again. My problem this time is that I always feel that when I push the stick a small distance, it is faster than pushing the same distance in the game using controller directly.So I simply did some tests. I used rewasd in Tarkov and compared it with Battlefield 1 and Apex. Both games use a linear curve, and tarkov with rewasd also use linear curve.But it is difficult for me to control the precise numerical experiments manually, so I only did rough experiments. This proves that when the maximum speed of the three games is adjusted to the same(When the stick is pushed to the maximum, the speed in the game is the same), the linear simulation of rewasd is indeed faster when the input is small. And when the dead zone is set to 8%（it is 2621 in rewasd）, rewasd gives me a completely different feeling from the other two games, and it feels much smaller. Based on these two points, I make the following guess. The game that directly uses the joystick is based on the distance pushed by the finger, while the joystick input used by rewasd is based on the angle, because when the joystick is pushed to a small degree, it may rotate 1 degree by pushing 1mm, but when it is fast When pushing to the end, you may need to push 3mm to rotate 1 degree, and there is a tangent conversion between angle and distance. This also explains why rewasd gets faster results when pushing the joystick slightly. For games that natively support controllers, the input signal has been converted, so push the stick slightly more slowly. But these are just my guesses. Can you use Battlefield 1 or Apex(linear curve) to do some accurate numerical tests for comparison based on this theory? If you can, can you try to adjust the input method to the same as these games? Thank you very much!

Hello!

You are using a mouse mapped to a stick, right? The games treat those inputs differently, so this is why the input is not the same and it can't be 100% the same out of the box. However, with Response curve, you can adjust it the way you feel it should be for the certain game. As you may see in this thread, all the curves from other folks have the lowered first point, and this is the thing that will override the issue you are describing.

Mouse emulation is a universal thing, it can't be based on one or a few games. But with a Response curve, any customization is possible.

[ventus]

No, you did not understand what I meant.I used the exact same curve as in APEX, but I even think the result is a convex curve.
This is the curve in APEX.

This is the curve I used in rewasd.

But it is still much faster than the feeling when I use a linear response curve in APEX.I even think the result is a convex curve.
After my test, all games that natively support controllers should process the stick's raw input signal through tangent transformation and convert it into distance, and then different games have different curves to adjust this converted signal. I also tried the curves of the above several people, and even I pulled the nodes to a very low level, but he was still fast, just one custom curve is not enough. Because this is actually the effect of the superposition of two curves, if you want to combine it into a function, even 16 points are not enough. And These changes will not affect the UI at all, only an extremely simple adjustment in the algorithm. If have the source code, I think the change can be completed in less than five minutes, can you just try it?You don’t have to release it, just give me a test version？

[reWASDer]

We are planning to add other types of curves that won't be related to points, so I hope it may do the trick.

Could you please send me the function for the curve you want to get?

However, I can't promise you any fast release dates, it won't be done in the upcoming version.

[ventus]

We are planning to add other types of curves that won't be related to points, so I hope it may do the trick.

Could you please send me the function for the curve you want to get?

However, I can't promise you any fast release dates, it won't be done in the upcoming version.

This is the fitting curve I got after testing Battlefield 1 and Battlefield 5. But because it is based on the scattered points, the result is a polynomial, which is a bit strange, and there should be a better solution than the polynomial.

[reWASDer]

Good, thanks for sharing!