well I try to do

the same very simple math in actionscript and puredata and get different results...

PD

1.5707963267948966 -> cos -> print

output: -4.37114e-008

in flash

trace(Math.cos(1.5707963267948966));

output : 6.123031769111886e-17

I think it's because the number is cut when too long or something,

is there a way to fix this ?

thanks

http://lists.puredata.info/pipermail/pd-list/2012-04/095889.html

does that mean I'm screwd ?

This isn't directly related to your link. The reason for this is because Pd uses the single-precision 32-bit floating-point data type for all numbers. It looks like ActionScript is interpreting it as a double-precision 64-bit floating-point.

Pd isn't coming up with the wrong value; it's as accurate as you'll get with 32-bit floats. In most cases, it isn't a problem (both of those numbers are really, really small), so I wouldn't necessarily say you're "screwed".

Hey !

thabnks for your answer, but is there a way to easy make bigger precision ?

I'm trying to make a patch against gimbal lock for 3D rotation from kinect translations.

so I have my XYZ angles, I convert them to a matrice, check condition for gimbal lock and convert back to degree,

but while all of this there is a lot of cos sin & tan added and multiplied wich create very different results from for example flash and the part I'm intrested in for gimbal lock fix are very tiny precise numbers ...

so ? any lutions ?

thanks baby cot ! very helpfull

though maelstorm I now realise thant when turned back in degree it fits

thanks cheers

I have the same proble with rotating 3d and gimbal lock solution in pd. Did you finished your patch ?
it will be great if you can share