comp.lang.idl-pvwave archive: archive

Home » Public Forums » archive » accuracy problems

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: Accuracy problem [message #47938 is a reply to message #19921]

Sat, 11 March 2006 22:45

kehcheng
Messages: 2
Registered: May 2004

Junior Member

A better solution is to use the formula ATAN( NORM(C), TRANSPOSE(A)#B),
where C is A cross B, for the angle between vectors A and B. There is
no need to limit the two arguments of ATAN, and you'll get a much more
accurate result when A and B are almost parallel (or antiparallel).
Example: A=[1.d0,0,0], B=[1.d0,0,1.d-9]. The ACOS formula gives 0.0
for the angle; the ATAN formula gives 1.d-9.

Report message to a moderator

[Message index]

		accuracy problems By: J�rg Schliwa on Thu, 27 April 2000 00:00
		Re: Accuracy problem By: kehcheng on Sat, 11 March 2006 22:45
		Re: Accuracy problem By: Juan Arrieta on Sat, 11 March 2006 19:37
		Re: Accuracy problem By: mmeron on Sat, 11 March 2006 16:25
		Re: Accuracy problem By: David Fanning on Sat, 11 March 2006 06:14
		Re: Accuracy problem By: James Kuyper on Sat, 11 March 2006 03:56
		Re: Accuracy problem By: mmeron on Fri, 10 March 2006 22:52
		Re: Accuracy problem By: David Fanning on Fri, 10 March 2006 21:43

Previous Topic:	Re: How can I find the central pixel?
Next Topic:	Floats

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Tue Dec 02 23:19:58 PST 2025

Total time taken to generate the page: 2.15943 seconds