How can I zoom out an image then rotate it 30 degrees by its center, then flip vertically the rotated image, while keeping it rotated?
Actually the solution was quite simple.
It's not about math - i.e. mixing the transformation matrix - it's about js and the way canvas works.
I've actually forgot to call context.save() and context.restore() before and after (all) my transformations. So actually half of the transformations were applied, and the other half were applied on the next iteration.
So now I just do context.save , then all my transformations, then context.restore at the end. Note that transformations are cumulative, i.e. one rotation of 1 degree followed by another rotation of 1 degree results in a 2 degree rotation at the end.
Related
My goal is to bring in an image on to part of the canvas, then scale it, move/translate it, and optionally skew it, also rotate and make alpha changes, kind of the primary "2d image manipulations", in an animated form, which is: do little changes over time from the starting state to the target end state.
Well, I figured to be efficient, I should use the canvas/2d context transform, https://developer.mozilla.org/en-US/docs/Web/API/CanvasRenderingContext2D/transform -- as it does the first 3: scale, move/translate, and skew, "all in one." I did half that code, and now I'm looking at examples and seeking to debug it. All the examples I see, are do 1) some transform, away from the "unity transform":
{ a:1, b:0, c: 0, d:1, e:0, f:0 }; // this basic transform does nothing
and then 2) draw into that. But that's the opposite order from what I want: which is draw on the canvas (the image), and then do an animation over time using the above primary changes (scale, translate, skew, rotate, and alpha). My question is: does it only "work this way", meaning I must setup the (single) transformation on the page first, and then "draw into that?"
I hope not ... that won't give me what I want, and I have to "ditch it", and go to 5 individual "transformations." Comments?
Yes that only works this way, canvas transforms and compositing mode and filters and lineWidth and fillStyle etc. properties are only applied to the next drawing operations.
The canvas itself only holds pixels information, it has no concept of drawn object. Your js code has to do this part.
So for what you wish, you can simply redraw everything every time:
reset the transform so we can clear ctx.setTransform(1,0,0,1,0,0);
Clear the canvas ctx.clearRect(0,0,ctx.canvas.width,ctx.canvas.height)
Set the transform to your new matrix ctx.translate(x,y); ctx.scale(s)...
Draw your transformed graphics ctx.fill(); ctx.drawImage(...
Wait next frame to do it again requestAnimationFrame(update)
I try to use 2 ellipses to detect a collision if they overlap. I have to rotate the ellipses but I can't figure out how this works. I'm working with the "com.badlogic.gdx.math.Ellipse" class but it seems to have no method for rotating. Any ideas? Thx in advance!!
Unfortunately, LibGDX doesn't have in-built rotating functions for ellipses.
Instead, I'd either be resorting to a circle in which rotation does not matter, or use polygons to check intersection.
Polygons are formed through an array of float values (vertices), where, every even element of the array is the horizontal component (x) and the odd, the vertical component (y).
Polygon polygon1 = new Polygon(vertexSet1);
Polygon polygon2 = new Polygon(vertexSet2);
Then, by using an Intersector, you can then check whether these polygons have intersected. The more vertices, the more accurate your shape will be. Just remember to have 6 or more elements in your vertex array, as the 6 floats will give 3 (x, y) points which is the minimum required for a polygon.
if (intersector.overlapConvexPolygons(polygon1, polygon2) {
//do your intersection code
}
The polygons themselves have commands to translate, scale and rotate, allowing for the rotations you mentioned above.
I'm not very good with radial calculations, I can't imagine thus I can't be sure. I need some explanation of Math.atan2() thing, please.
Usual task - to make an object rotate after the mouse. I get the differences, get the angle, I see angles in the text areas and DIRECTLY the object does follow the mouse. What I need now is everything to be smooth. I need angles to be 0-360 but after 180 object rotation becomes -180 and counts backwards, and mouse rotation becomes -90 after 270 and also counts back to 0.
More deeply, I want a smooth rotation, it means a set speed of say 2 per frame, to meet the mouse angle the shortest way. It takes to set conditions and I can't do that cause I don't even understand the logic of these values. They are almost random! I don't need it to be done or copied, I need to understand to move on so if you could please explain how does it work and what I do wrong...
Code is simple:
angle = Math.atan2(deltaY,deltaX)/(Math.PI/180) + 90; //+90 cause it lacks it to look at the mouse//
Object01.rotation = angle;
So the problem is I don't even get how it works... if 2 values are different the object can't point at the mouse but it does. Numbers lie and if I need something based on these numbers it will be wrong. Very wrong... Need organization. Meaning I want everything to be ready for further coding that will be based on the rotations to not jump up and down cause of misfit ends.
Add: Explanation of how does it happen, what I described above. Why such a chaos of the values? And an advice on how could I arrange it for further coding, just as I said. Animation alone wont work if I want to make rotation an element of important events such as shooting direction and aiming speed. Or changes of speed rotation of a lockpicked lock. Or anything much more complicated that wont work if I don't make straight and clear values: from A to Z, from 1 to 10, no 8s between 2 and 3, no R before B, no mouse angle 270 when object facing it -90 when they both started from 0 and reached 180 together.
Oh, and as I said, mouse facing works but when I try to make a certain speed of chasing mouse the shortest way it turns the object wrong directions in all 4 quarters. I assume it's also about this arctangens thing that has issues with delta values becoming negative in different quarters. And when I change it, some other value goes wrong... So I need to know exactly what I'm doing to know what's wrong and how to fix it. So yep. Need explanation. Please.
Add2: angleZ = Math.atan2(oppSide,adjSide)/(Math.PI/180);
So I divided rotation to 4 quarters, for each I count atan as opp. side to adj. side, then add 90, 180 and 270 respectively. My mouse rotation does 360, but the object that follow through simple object.rotation = angleZ; still goes to 180, then from -180 to 0 on the left side. Why does it ignore the simple command? The rotation fits but I need it to be equal, no surprises! Why is it happening? How can a number I directly set to be equal to another number as a base of the action change itself to the one of same rotation but completely different number? It doesn't even know it's degrees! It's as simple as "object.rotation, please be equal to the number I choose!"
It's just different coordinate systems. Like how x starts at 0 at the left of the stage, goes +x to the right, and -x to the left, object rotation starts at 0˚ pointing up, and goes +180˚ clockwise and -180˚ anti-clockwise.
Math.atan2 happens to start at 0 pointing left (-x), and go +270˚ clockwise and -90˚ anti-clockwise, which is annoying, but it just means you have to convert between coordinate systems by adding 90˚.
You can spin something around over and over of course, so the numbers jump so that they always stay within the same range, because 361˚ is the same as 1˚, and -270˚ is the same as 90˚. You can tell an object to rotate outside of the -180˚ to 180˚ range, and it will normalise the rotation to within those values.
As mitim described, to smoothly animate rotation you'll either need to use Event.ENTER_FRAME, a Timer, or a tweening library like TweenLite. Most tweening libraries can work out the shortest rotation direction for you, but otherwise you can simply calculate both and see which is smaller.
As an idea, since it seems like you know the angle you need to rotate towards and it's direction, would it be easier to just animate towards that angle to get your smooth rotation? Basically treat it like any other animatable property and just add on your rotation speed (2 degrees it looks like) per frame tick, until it reaches the desired rotation.
Find angle amount needed to rotate by
Figure out if clockwise or counter clockwise direction and set the rotation amount. This can be figured out by checking if the angle is great then 180 / positive or negative
Add the rotation amount * direction every frame tick, until the desired rotation is less then or equal to the rotation amount per frame
Set rotation to desired rotation
This is my firs excursion on the HTML5 canvas, I have working knowledge of jQuery and Javascript.
I'm trying to create a "spinning globe" effect with it.
The idea is to have a circle and meridians "spinning" on it, to give the effect of a rotating globe.
I've drawn the circle and now I'm trying to create lines that start from the right (following the curve of the circle), move towards the centre straightnening up (in the middle they are straight) and follow the inverse curvature on the left, ending with the circle.
I'm trying to do this with the HTML5 canvas and jQuery but I'm not sure of where to start... should I create an arc and then try to animate it?
I'm even wondering if the canvas is the right tool or if I should use anything else.
Any suggestion is welcome!
Sebastian
You could use a quadratic bezier curve, which is basically just a curve with a start point, an end point, and a "control point" in the middle, which is what you would want to change as the globe spins. In this case, all of your lines would start and end at the north and south poles, respectively, of your "globe". For example, to make one of these lines:
// start drawing a line
canvas.beginPath();
// move the the top of your globe
canvas.moveTo(0,0);
/* draw a curve with the control point specified by the first two args,
* end point by the second two:
* (in your case, the control point would be in the middle of the globe) */
canvas.quadraticCurveTo(control_point_x, control_point_y, 0, 50);
// finish drawing, stroke and end
canvas.stroke();
canvas.closePath();
You would also have to take in to account how you will clear the lines after each frame, of course.
See: The Canvas element API, Complex Paths
This is what I got, didn't have the time to proceed any further: http://jsfiddle.net/Z6h3Z/
I use bezier curves where the two control points are in a sort of oval arc centered at the poles.
What I got stuck at is the distribution of points along the arc to look more realistic.
In Flash, pixels are calculated using twips, or twentieth of a pixel. Consequently, every position is always in multiples of 0.05. I haven't seen this mentioned in the HTML Canvas spec and am unable to trace the cursor position on Canvas. Does anyone know the accuracy of its pixel calculations?
Edit for clarification:
I'm referring more to Zeno's paradox which says in order to move something from point A to point B, it must first move to a point halfway between the two. And then halfway again, ad infinitum.
So if I want to move on the x axis from point 0 to 100 at 0.5:
At frame 1: 50
Frame 2: 75
Frame 3: 87.5
Then: 93.75, 96.875, 98.4375... etc.
So at what step does the Canvas actually round-up to the next pixel?
I'm unsure what you mean by accuracy of slicing.
Pixels on the Canvas can be drawn to a little less than 0.10, after which they make barely any visible impact.
Of course, if you scale, you can draw things that are 0.00125 pixels, and so on. But they won't be visible if you unscale.
http://jsfiddle.net/GvVD9/
(That first square block on the top-left is a pixel)
Accuracy of the mouse is an entirely different thing, in no way related to the canvas spec.
EDIT:
Well, we can sorta demonstrate that. We can draw a bunch of pixels with y values approaching 100 and see how they compare to a red pixel drawn with the y value 100.
http://jsfiddle.net/GvVD9/46/
Every single horizontally separated piece is just a single 1 by 1 pixel rect using the drawRect command.
50
75
87.5
93.75 // first black pixel you see in image
96.875
98.4375
99.21875
99.609375
99.8046875
99.90234375
99.951171875
99.9755859375
99.98779296875
99.993896484375
99.9969482421875 // last black pixel you see in image