Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84; site lanl.ARPA Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!godot!harvard!seismo!cmcl2!lanl!jlg From: jlg@lanl.ARPA Newsgroups: net.audio Subject: Digital filter Message-ID: <17126@lanl.ARPA> Date: Fri, 30-Nov-84 20:24:09 EST Article-I.D.: lanl.17126 Posted: Fri Nov 30 20:24:09 1984 Date-Received: Sun, 2-Dec-84 05:39:27 EST Sender: newsreader@lanl.ARPA Distribution: net Organization: Los Alamos National Laboratory Lines: 19 > ( line eater, line eater, eat me a line. Oh, while your at it, one for me too) > > " A digital transversal filter is used for the filtering after > oversampling. To understand the operation of the filter, we can > think of it as consisting of 96 elements, while the delay in each > element is (176.4 * 1000) ** -1s, i.e. a quarter of the sampling > period or 1/4 Ts. Four times in each period the filter takes up new > data. At three of these four times, the content of this data is > zero, since the oversampling is done by the introduction of > intermediate samples of value zero. This means that only 24 of the > 96 elements are filled at any one time. The contents of each > element are multiplied by a coefficient c. The filter provides data > at a rate of 176.4 kHz; each number is the sum of 24 non-zero > multiplications. In this way the filter always calculates three new > sample values at the locations of the zero samples." In other words, it's a 96 element discrete convolution integral implemented in hardware. Just what you'd expect a digital filter to look like.