KeatonMill

10th December 2002 00:54 UTC

Wrapping Variables
Hey guys. I've got a question. Is there a way to wrap a variable in superscopes? Specifically, I want to wrap something from -1 to 1, so that if the value is say 1.5, the wrapper will change it to -.5. I've tried different combinations of trig and arctrig functions, but haven't gotten anything to work yet.

Can someone help?

Thanks,
Keaton

UnConeD

10th December 2002 01:26 UTC

You can use a hack with the % operator to wrap with a certain precision... for example to wrap 'a' between 0 and 5 to a precision of 0.01:

a=((a*100)%500)*.01

To wrap between -5 and 5 with a precision of 0.1:

a=(((a+5)*10)%100)*.1-5

(add 5, scale by 10, wrap between 0-100, scale by 0.1, subtract 5)

This isn't perfect and has some troubles near zero, but it works for most purposes.

This technique will cause the numbers to jump back to the lower limit when they reach the upper limit.

There's a different technique which has the numbers bounce between the upperlimit and lowerlimit (ping-pong style):

asin(sin(var*(pi/2)))/(pi/2)

This resembles a sine-wave, except it's completely linear. Naturally you should replace pi/2 and 1/(pi/2) by constants, but I left them in for clarity.

Jaheckelsafar

10th December 2002 06:33 UTC

You could go the simple rout.

a=a+.1;
a=if(above(a,1),a-2,a);
a=if(below(a,-1),a+2,a);

It might no be the fastest, but it works. I have proof.

UnConeD

10th December 2002 07:23 UTC

Jaheckelsafar, your method works for variables that increase/decrease less than one period of the wrapping. Mine works for wrapping any number into the range.

So, choose whichever one you need. Jaheckelsafar's is faster of course.

fsk

13th December 2002 11:08 UTC

this might sound silly but what is wraping?

nixa

13th December 2002 18:53 UTC

FSK: Wraping means to get any variable value to be between some min and max values. Moastly this is used whith angles alpha=alpha+2*pi=alpha+4*pi=alpha+6*pi=alpha+2*k*pi,(k element of Z)
so if you have some angle alpha=7pi you will want to write it as alpha=pi

jheriko

14th December 2002 17:03 UTC

if you want to use trig functions for wrapping then you could use atan(tan(poo)); and it will wrap poo between about -1.6 and 1.6

dirkdeftly

14th December 2002 23:50 UTC

asin(sin(x*pi*a)/pi/a) wraps from -a to a. If it's not working try asin(sin(x*pi/a)/pi*a).