Comments on: Squaring The Thumbsticks http://theinstructionlimit.com/?p=531 Truevision3D 6.5, XNA, C# and HLSL tips & samples by Zaknafein/Renaud Bédard. Thu, 02 Sep 2010 18:01:19 -0700 hourly 1 http://wordpress.org/?v=3.0.1 By: Jack http://theinstructionlimit.com/?p=531&cpage=1#comment-274 Jack Fri, 13 Nov 2009 02:56:09 +0000 http://theinstructionlimit.com/?p=531#comment-274 Congratulation Renaud!:) Congratulation Renaud!:)

]]> By: raigan http://theinstructionlimit.com/?p=531&cpage=1#comment-265 raigan Wed, 04 Nov 2009 21:19:42 +0000 http://theinstructionlimit.com/?p=531#comment-265 Yeah, actually I don't think what I was thinking of could work, I just tried to write it out and it would still produce circular patterns :( Nice that you found a solution though! Yeah, actually I don’t think what I was thinking of could work, I just tried to write it out and it would still produce circular patterns :(

Nice that you found a solution though!

]]> By: Renaud Bédard http://theinstructionlimit.com/?p=531&cpage=1#comment-264 Renaud Bédard Wed, 04 Nov 2009 17:42:59 +0000 http://theinstructionlimit.com/?p=531#comment-264 I'm still not sure what you meant, because... the input values are inside a circle, I have to scale them by something to get the square values. And interpolation is still an approximation, based on what I tried. (unless you mean some other kind of interpolation) But anyway, I figured out what I think is a perfect and flexible solution, see the update notes! I’m still not sure what you meant, because… the input values are inside a circle, I have to scale them by something to get the square values. And interpolation is still an approximation, based on what I tried. (unless you mean some other kind of interpolation)

But anyway, I figured out what I think is a perfect and flexible solution, see the update notes!

]]> By: raigan http://theinstructionlimit.com/?p=531&cpage=1#comment-262 raigan Sun, 01 Nov 2009 22:22:08 +0000 http://theinstructionlimit.com/?p=531#comment-262 ..basically since you have a perfect square, it should be possible to produce and output that's a perfect square gradient, instead of that clipped-circular gradient that's currently being produced. One way to do that would be to bilinearly interpolate between (1,1),(1,-1),(-1,-1),(-1,1) (i.e the four corners of the square) based on the current input position within the square. I hope this makes sense! ..basically since you have a perfect square, it should be possible to produce and output that’s a perfect square gradient, instead of that clipped-circular gradient that’s currently being produced.

One way to do that would be to bilinearly interpolate between (1,1),(1,-1),(-1,-1),(-1,1) (i.e the four corners of the square) based on the current input position within the square.

I hope this makes sense!

]]> By: raigan http://theinstructionlimit.com/?p=531&cpage=1#comment-260 raigan Sun, 01 Nov 2009 14:08:28 +0000 http://theinstructionlimit.com/?p=531#comment-260 Hmm.. I may have misunderstood the goal of this exercise :) Can I modify my suggestion? Once you clamp the input to within the inscribed square, you'd then want to interpret the input in the context of the square. I don't really know what that last part means, I guess it would just be calculating the distance from the input position to the square's edges -- it should be just like bilinear interpolation. Hmm.. I may have misunderstood the goal of this exercise :)

Can I modify my suggestion? Once you clamp the input to within the inscribed square, you’d then want to interpret the input in the context of the square. I don’t really know what that last part means, I guess it would just be calculating the distance from the input position to the square’s edges — it should be just like bilinear interpolation.

]]> By: raigan http://theinstructionlimit.com/?p=531&cpage=1#comment-256 raigan Tue, 27 Oct 2009 01:49:22 +0000 http://theinstructionlimit.com/?p=531#comment-256 Yeah, that sort of thing really bugs me.. stupid XNA :) Yeah, that sort of thing really bugs me.. stupid XNA :)

]]> By: Renaud Bédard http://theinstructionlimit.com/?p=531&cpage=1#comment-255 Renaud Bédard Mon, 26 Oct 2009 17:12:43 +0000 http://theinstructionlimit.com/?p=531#comment-255 Hehe, MathHelper isn't my class, it's part of the XNA framework. I actually have a MathHelperEx.Sqrt2 in my own libraries, but one may choose to just use a local const field or whatever. Hehe, MathHelper isn’t my class, it’s part of the XNA framework. I actually have a MathHelperEx.Sqrt2 in my own libraries, but one may choose to just use a local const field or whatever.

]]> By: raigan http://theinstructionlimit.com/?p=531&cpage=1#comment-254 raigan Mon, 26 Oct 2009 17:06:05 +0000 http://theinstructionlimit.com/?p=531#comment-254 N+ XBLA uses an even simpler method -- no analog control, the thumbsticks are just thresholded to 0 or 1. One thing is bugging me about the code: you have MathHelper.PiOver2 but no MathHelper.SqrtOf2 :p N+ XBLA uses an even simpler method — no analog control, the thumbsticks are just thresholded to 0 or 1.

One thing is bugging me about the code: you have MathHelper.PiOver2 but no MathHelper.SqrtOf2 :p

]]> By: Renaud Bédard http://theinstructionlimit.com/?p=531&cpage=1#comment-251 Renaud Bédard Mon, 26 Oct 2009 02:26:46 +0000 http://theinstructionlimit.com/?p=531#comment-251 That's just boring ;) But it's probably fine for thumbsticks, yup. Good to know. That’s just boring ;)
But it’s probably fine for thumbsticks, yup. Good to know.

]]> By: raigan http://theinstructionlimit.com/?p=531&cpage=1#comment-250 raigan Sun, 25 Oct 2009 23:17:29 +0000 http://theinstructionlimit.com/?p=531#comment-250 Or there's the lazy solution: inscribe a square in the circle, anything outside of the square gets clamped to the surface of the square (i.e to max value). Good enough :) Or there’s the lazy solution: inscribe a square in the circle, anything outside of the square gets clamped to the surface of the square (i.e to max value). Good enough :)

]]>