Cartogrammar·Blog

[...] ensured isolines would still pass through the interpolated values. You can read about the method in his post. Here she [...]

indiemaps.com/blog » isolining package for ActionScript 3
9 June 2008 @ 1:16am

[...] Continuous curves with ActionScript 3 [...]

Kelso’s Corner » Blog Archive » New blog: Cartogrammer
15 June 2008 @ 2:46pm

Hi…
Nice work…considering your ‘limited understanding of the math involved’ … Q: Am I blind? Is that i don’t see where you finally use the z factor (third argument of the curveThroughPoints method)… Thanks

alvaro obyrne
20 June 2008 @ 2:37pm

Yikes, thanks for noticing that! Looks like I hard coded a number in when I was messing around with it and forgot to change it back.

Line 146 of CubicBezier.as, which says:
var controlDist = Math.min(a,b)*.5
Should be:
var controlDist = Math.min(a,b)*z

I’ve loaded the corrected file to the links given in the post.

Andy Woodruff
20 June 2008 @ 2:55pm

thanks

alvaro obyrne
20 June 2008 @ 3:07pm

Sorry if you don’t consider this is not the place to post this but have you seen the java demo at http://www.sunsite.ubc.ca/LivingMathematics/V001N01/UBCExamples/Bezier/bezier.html the athor of that applet doesn’t use 100 points in each segment of the curve. He uses a variable number of steps according to the curvature (the straighter the less number of points he uses). If I may say: that should be your next step (I should even farther say : this should be our next step).

alvaro obyrne
20 June 2008 @ 3:19pm

Thanks for the link; I hadn’t seen that. Yeah, it’d be good to draw the curves like that. I don’t know that I’ll get around to thinking about that very soon, but if you feel like trying it out I’d definitely love to see what you come up with!

Andy Woodruff
20 June 2008 @ 6:55pm

On line 150
var theta = Math.atan(ry / rx);

if ry and rx is 0, theata gets NaN, because it cant be 0/0

if (isNaN(theta)){
theta = 0;
}

Is it right? =)

//Dennis Öhman =)

Dennis Öhman
4 August 2008 @ 4:53am

Sorry, it’s not right solution =D

Dennis Öhman
4 August 2008 @ 9:41am

Hi there! Your class is awesome! I did however notice that it messes up if the line between two points is exactly straight. I made a fix by adding a Boolean variable called isStraight that tracks whether it is a straight line or not, and then at the end drawing a straight line instead of curves in that instance.

Add the below line of code right after the a,b,c distance calculations:
isStraight = ((a+b)(c-1)) ? true : false;

and add similar to:
if (!isStraight) {
g.curveTo(controlPts[1][0].x,controlPts[1][0].y,p[1].x,p[1].y);
} else {
g.lineTo(p[1].x,p[1].y);
}

in the three sections at the bottom. Feel free to email me for the code itself if this is not clear. Thanks again for the great work!

John Currie
5 September 2008 @ 5:31pm

oops, the first line of code there should be:
isStraight = ((a+b)<(c+1)&&(a+b)6>(c-1)) ? true : false;

John Currie
5 September 2008 @ 5:32pm

minus the ‘6′ there sorry!

John Currie
5 September 2008 @ 5:32pm

Thanks! I sort of recall noticing that problem a while ago but either never got around to trying to fix it or never got around to uploading a new version. (Same goes for comment number 8.) When I get a chance I’ll be sure to update the posted file with that fix.

Andy Woodruff
6 September 2008 @ 9:13pm

I added a moveto:Boolean parameter with default of false, and then added a conditional to the moveTo in the curveThroughPoints method. This allows you to join up several curveThroughPoints into a single fillable object that has cusps as well as vertices if I am using the terminology correctly.

slaingod
12 September 2008 @ 12:24pm

Try send these values to the class

(x=0, y=0)
(x=70, y=1)
(x=141, y=3)
(x=211, y=4)
(x=282, y=6)
(x=352, y=2)
(x=423, y=2)
(x=493, y=2)
(x=564, y=18)
(x=634, y=60)

If there is 3 equal values, like Y=2 Y=2 Y=2.
The last one of the equal numbers becomes x=0 y=0 (Or something?).

I don’t know how?

Dennis Öhman
12 November 2008 @ 3:55am

For example:
If you line upp the dots in you example, this happens

http://devarea.webbhotell.tv/fb/temp/crap.png

Dennis Öhman
12 November 2008 @ 4:01am

It’s med again hehe
Why do you do this?:

// Loop through duplicates array and remove points from the points array
for (i=duplicates.length-1; i>=0; i–){
p.splice(duplicates[i],1);
}

Dennis Öhman
12 November 2008 @ 4:09am

Sorry, I should read in the class before I post a question =D

17. I know what i t does.

Dennis Öhman
12 November 2008 @ 4:18am

Hm. I can’t seem to figure out how to get around the straight-line problem. John’s fix isn’t working for me. Where exactly are you definining the isStraight bool? It would be great to get this into the class itself. Otherwise excellent work! Very useful code!

George Berstrand
14 November 2008 @ 3:40am

Use atan2() instead. I’ve tried it and it works
And for the acos() in the source code, I transformed the cosine expression in to a tangent expression and then used atan2(), correctness also verified.

CJ Cat
15 November 2008 @ 9:35am

This is an awesome class!

However, has anyone implemented an AS2 version? Would be greatly appreciated…

Cheers
Matt Stuehler

Matt
15 November 2008 @ 1:38pm

Andy,

Have you had any time to look at the straight line problem? I’ve tried the other suggestions in the comments here with no success. CJ - the atan2 method seems to produce an s-curve instead of a straight line. Aside from that little glitch this is exactly what I was looking for.

Thanks!
Sam

Sam
7 December 2008 @ 5:36pm

Sam: Unfortunately I haven’t had a chance to look at much of anything on this class lately. Apologies to those encountering problems (but I’m glad for the interest)! It’ll be a week or so until I have time, but I’ll do my best to study it then and see if I can think of anything.

Andy

Andy Woodruff
8 December 2008 @ 10:05am

[...] me di cuenta que lo que necesitaba obetener se basaba en la interpolación de curvas bézier y aquí encontré una buena solución basada en curvas bézier contínuas. Lo curiose es que siempre andan [...]

Bubbles V1.0 - Continuous bezier curves | arturoparacuellos blog
13 January 2009 @ 1:25am

TypeError: Error #1010: at curveTest_fla::MainTimeline/moveDot()

SmivaL
20 June 2009 @ 12:47pm