文件列表:

remez (0, 2005-07-17)
remez.pas (20001, 2005-07-16)
remez\build.bat (213, 1999-07-18)
remez\Debug (0, 2004-06-17)
remez\filter.c (1732, 1999-07-18)
remez\filter.dat (902, 2003-11-30)
remez\GPL (18342, 1999-07-18)
remez\lex.yy.cpp (43051, 1999-07-18)
remez\LGPL (25747, 1999-07-18)
remez\lowpass.bat (39, 1999-07-18)
remez\lowpass.csv (23718, 1999-07-18)
remez\lowpass.m (7146, 1999-07-18)
remez\lowpass.rem (182, 1999-07-10)
remez\main.cpp (6846, 1999-07-18)
remez\remez.cpp (20144, 1999-07-18)
remez\remez.exe (73728, 1999-07-18)
remez\remez.h (1834, 1999-07-18)
remez\remez.l (2835, 1999-07-18)
remez\remez.tab.cpp (25070, 1999-07-18)
remez\remez.tab.h (290, 1999-07-18)
remez\remez.y (2387, 1999-07-18)
remez\t2d.csv (23649, 2003-11-30)
remez\t2d.m (1398, 2003-11-30)
remez\t2d.rem (177, 2003-11-30)
remez\t2d.rem.bak (181, 2003-11-30)
remez\test.bat (30, 1999-07-18)
remez\test.csv (23582, 1999-07-18)
remez\test.m (1293, 1999-07-18)
remez\test.rem (532, 1999-07-18)
remez\types.h (1184, 1999-07-18)

------------------------------------------------------------------------- README file for Remez FIR Filter Design Program Copyright (C) 1999 Jim Ledin (jim@ledin.com) This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA ------------------------------------------------------------------------- This program (remez.exe) designs optimal linear phase FIR filters using filter specifications read from an input file. It is based on an implementation of the Parks-McClellan equiripple design program by Jake Janovetz (janovetz@uiuc.edu). Input file format: Statements in the input file can be arranged in free format. There should be at most one of each type of statement {order, bands, amplitudes}. {type, weights} are optional statements with default values as specified below. Numerical values must be in C floating point format except that no exponents or minus signs are allowed. Mandatory statements: order = integer The filter order. The number of filter coefficients = order+1. bands = (number, number) [, (number, number) ...] Each pair within parentheses contains the lower and upper band edges, with lower listed first. The frequency is defined in the range [0,1] where 1 is the Nyquist frequency, which is half the sampling rate. Bands must be listed in increasing frequency order. The first band must begin at 0 and the last band must end at 1. The separation between bands must be greater than zero to allow for a transition region. amplitudes = number [, number ...] Each amplitude goes with the corresponding band in the band statement. Amplitudes must be >= 0. Optional statements: type = bandpass | hilbert | differentiator defaults to bandpass if this statement is not provided. weights = number [, number ...] Each weight goes with the corresponding band in the bands statement. Weights must be greater than 0. All weights default to 1 if this statement is not provided. Comments: Comments can be included in the input file using C/C++ syntax: // This is a C++ end-of-line comment /* This is C-style comment */ Example input file (test.rem): // Lowpass filter example // Note: type defaults to bandpass & weights default to 1 // This file describes a lowpass filter with a desired amplitude // of 1 in the frequency range [0, 0.45] and zero in the range // [0.55, 1]. The range [0.45, 0.55] is a transition region. // For an 8 KHz sampling rate the Nyquist frequency is 4 Khz and // the passband will be 0 to 1800 Hertz and the stopband will be // 2200 to 4000 Hertz. type = bandpass weights = 1, 1 order = 22 bands = (0, 0.45), (0.55, 1) amplitudes = 1, 0 Execute the command file test.bat to process this file. The filter.dat output file produced by this example is: #define N 23 static double H[N] = { -3.279199345817378200e-002, -9.447491858515190300e-005, 2.540825979405659000e-002, -7.219108010618943700e-005, -3.720689102197431700e-002, -6.954083711178532800e-005, 5.770408717695533300e-002, -8.033893***8899294400e-005, -1.023035371036797900e-001, -9.592081101169122500e-005, 3.171424233766678100e-001, 4.999104052948560900e-001, 3.171424233766678100e-001, -9.592081101169122500e-005, -1.023035371036797900e-001, -8.033893***8899294400e-005, 5.770408717695533300e-002, -6.954083711178532800e-005, -3.720689102197431700e-002, -7.219108010618943700e-005, 2.540825979405659000e-002, -9.447491858515190300e-005, -3.279199345817378200e-002 }; Program operation: usage: remez filename.rem [freqresp.csv] [filename.m] filename.rem = Input filter specifications freqresp.csv = Spreadsheet format frequency response data filename.m = Filter parameters in Matlab format C source output file = filter.dat The filter coefficients are always output to the file filter.dat. The file filter.c contains a functional, though not very efficient, FIR filter implementation that #includes filter.dat. Executing the command: remez lp.rem lp.csv lp.m will produce the output files filter.dat, lp.csv and lp.m. lp.csv contains frequency response data for plotting in a program such as Excel. lp.m is a Matlab script that defines the filter specifications for use with the Matlab remez function as: remez(N, F, A, W). It also contains the computed impulse response H which can be compared to the output of the remez function. Note that it is not necessary to generate the .csv or .m files. To just generate the filter.dat coefficients use the command: remez lp.rem Original program README file: /************************************************************************** * Parks-McClellan algorithm for FIR filter design (C version) *------------------------------------------------- * Copyright (c) 1995,19*** Jake Janovetz (janovetz@uiuc.edu) * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Library General Public * License as published by the Free Software Foundation; either * version 2 of the License, or (at your option) any later version. * * This library is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Library General Public License for more details. * * You should have received a copy of the GNU Library General Public * License along with this library; if not, write to the Free * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA * *************************************************************************/ DESCRIPTION: ============ The Parks-McClellan program is a filter design program which creates optimal filters in the sense that there is equirriple in the pass bands and stop bands of the resulting frequency response. This version was written for use in another program in which I required C access to the design routine. Therefore, Matlab's 'remez' function wouldn't do. I also didn't like all the FORTRAN to C conversions done because they're impossible to read. This version is based on the original FORTRAN code and equations presented in Oppenheim & Schafer. To use the code, simply include all but 'main' in your own code. There is no front end supplied with this code, but that should be simple to implement. If any errors are found, please let me know. I have compared the output to that of Matlab's 'remez' function and get close (within reasonable rounding error -- on the order of 1e-13) to the impulse response. However, I have not exhaustively tested it with a lot of vectors. Jake COMPILE INSTRUCTIONS: ===================== To compile the test program that generates a single filter, but asks for no input (it truly is a test program), very simply type: cc -o test test.c remez.c -lm On a UNIX machine. You may use other compilers (namely gcc), but you must link the math libraries (-lm). If you have problems compiling on a system, let me know and I'll see what I can do. The code should be very portable, though. Then, you can run the compiled program (test) and watch it generate some fascinating (!) coefficients. THANKS: ======= Thanks to Dr Peter Kootsookos for his help in finding a couple bugs in the original code.

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：