Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!doc.ic.ac.uk!uknet!pipex!uunet!news.cnri.reston.va.us!newsserver.jvnc.net!howland.reston.ans.net!ux1.cso.uiuc.edu!sdd.hp.com!col.hp.com!news.dtc.hp.com!hpscit.sc.hp.com!hplextra!hplntx!bhaskara
From: bhaskara@fc.hp.com (Vasudev Bhaskara)
Subject: MPEG audio
Summary: MPEG1 audio decode 
Sender: news@hplntx.hpl.hp.com (HPL News Posting Service)
Message-ID: <C931J8.3tH@hplntx.hpl.hp.com>
Date: Wed, 23 Jun 1993 16:23:31 GMT
Organization: Hewlett-Packard Laboratories, Palo Alto, CA
X-Newsreader: TIN [version 1.1.8 PL6]
Keywords: subband, layer 1
Lines: 20

In mpeg-1 audio standard, for layer1, they talk about a subband
analysis filter.
Basically, this operation involves multiplying a matrix N[i][k] by a
vector S[k] to get a vector U[i].

N[i][k] seems to be defined as...
    cos[(16 + i)(2k + 1)pi/64],  i = 0,1,..,63 and k = 0,1,..,31

If this matrix is defined as shown here, for i = 16, N[i][k] = 0 for
all k, which means U[i=16] = 0.
Is this definition of N[i][k] in error? (I think we should have it
has perhaps.. 16 + i + 1/2 instead of 16 + i.

Thanks for any pointers.
--
Vasudev Bhaskaran             Email: bhaskara@hplvab.hpl.hp.com
Multimedia Systems Deptt.     Phone: (415) 857-7153
Hewlett Packard Labs          Fax:   (415) 857-8526
1501 Page Mill Road
Palo Alto, CA 94304 
