From: Eric Durant <edurant@engin.umich.edu>
Subject: Re: Windowing [newbie question]
Date: 12 Mar 2000 00:00:00 GMT
Message-ID: <38CBDB8D.C8E5F826@engin.umich.edu>
Content-Transfer-Encoding: 7bit
References: <Pine.GS4.4.10.10003011334500.7196-100000@trurl>
<38c3bcb8.4735557@nntp.ix.netcom.com> <Pine.GS4.4.10.10003070814240.19969-100000@trurl>
<8a2s77$8jb$1@dino.active.ch> <8aais0$7ab$1@cowpat.doorknob>
X-Accept-Language: en,pdf
Content-Type: text/plain; charset=us-ascii
X-Complaints-To: news@eecs.umich.edu
X-Trace: news.eecs.umich.edu 952884171 79961 141.213.6.59 (12 Mar 2000 18:02:51 GMT)
Organization: University of Michigan Electrical Engineering & Computer Science Dept.
Mime-Version: 1.0
NNTP-Posting-Date: 12 Mar 2000 18:02:51 GMT
Newsgroups: comp.dsp

Gumboot wrote:
> Andor Bariska:
> > The Blackman window:
> > 0.42- 0.5 cos(2.0 pi / (N-1) k) + 0.08 cos (4.0 pi /(N-1) k)
> 
> That's something I can't get clear on; the value of N is...

Ah, the infamous off-by-one error and friends...   Harris, in his
classic paper on windowing, includes a clear discussion on common
indexing errors in windowing and apparent lacks of symmetry
(sec. II).  The Blackman window is also derived and its properties
are compared with a wide range of windows.

@INPROCEEDINGS{Har78,
   author = "Fredric J. Harris",
   title = "On the Use of Windows for Harmonic Analysis with the
            Discrete {F}ourier Transform",
   booktitle = "Proc. IEEE",
   volume = 68,
   pages = "51--83",
   month = jan,
   year = 1978,
}

-- Eric Durant
   http://www.edurant.com/