home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!ogicse!das-news.harvard.edu!das-news!kosowsky
- From: kosowsky@minerva.harvard.edu (Jeffrey J. Kosowsky)
- Newsgroups: comp.theory
- Subject: Re: Uniform noise in d-sphere
- Message-ID: <KOSOWSKY.92Nov9114504@minerva.harvard.edu>
- Date: 9 Nov 92 16:45:04 GMT
- Article-I.D.: minerva.KOSOWSKY.92Nov9114504
- References: <3655@news.cerf.net>
- Sender: usenet@das.harvard.edu (Network News)
- Followup-To: comp.theory
- Organization: Harvard Robotics Lab, Harvard University
- Lines: 23
- In-Reply-To: jcbhrb@nic.cerf.net's message of 6 Nov 92 02:25:25 GMT
-
- In article <3655@news.cerf.net> jcbhrb@nic.cerf.net (Jacob Hirbawi) writes:
- > A third method might be the following: use spherical coordinates and
- > pick uniform random numbers for each of the coordinates with the appropriate
- > ranges. In three dimensions this would be:
- >
- > (1) radius uniform over (0,d)
- > (2) angle1 uniform over (0,2 pi)
- > (3) angle2 uniform over (0, pi)
- >
- > This seems to be *too* simple but since I can't think of any point within the
- > sphere being more favored than any other point I would think that the
- > distribution is in fact uniform.
- >
-
- Not only is this *too* simple but it is also wrong. Picking radii and
- angles uniformly does *not* give a uniform distribution with respect
- to the standard (Euclidean) volume measure on the sphere. Remember the
- Jacobian term in the change of variables formula. Just because a map
- is one-to-one (eg: your spherical coordinate system is almost a
- one-to-one covering of the sphere) does not mean that it preserves
- measure!
-
- Jeff Kosowsky
-