- 安徽路人假
-
CIC抽取补偿滤波器设计,CIC滤波器采用5阶8倍抽取
如下:
%
%THIS IS A WIZARD GENERATED FILE. DO NOT EDIT THIS FILE!
%
%---------------------------------------------------------------------------------------------------------
%This is a filter withfixed coefficients
%This Model Only Support Single Channel Input Data.
%Please input:
%data vector: stimulation(1:n)
%
% This Model Only Support FIR_WIDTH to 51 Bits
%
%FILTER PARAMETER
%Input Data Type: Signed
%Input Data Width: 13
%FIR Width (Full Calculation Width Before Output Width Adjust) : 23
%-----------------------------------------------------------------------------------------------------------
%MegaWizard Scaled Coefficient Values
function output = CIC8_fir_comp_mlab_mat (stimulation, output)
coef_matrix=[0 0 1 0 -2 -3 0 7 10 1 -17 -26 -5 46 102 127 102 46 -5 -26 -17 1 10 7 0 -3 -2 0 1 0 0 ];
INTER_FACTOR = 1;
DECI_FACTOR = 1;
MSB_RM = 0;
MSB_TYPE = 0;
LSB_RM = 0;
LSB_TYPE = 0;
FIR_WIDTH = 23;
OUT_WIDTH = FIR_WIDTH - MSB_RM - LSB_RM ;
DATA_WIDTH = 13;
data_type= 1;
% check size of inputs.
[DX,DY] = size(stimulation);
[CX,CY] = size(coef_matrix);
if (CX ~= DY * INTER_FACTOR)
fprintf("WARNING : coef_matrix size and input data size is not match ");
end
%fill coef_matrix to length of data with the latest coef set
if (CX < DY * INTER_FACTOR)
for i= CX +1:DY * INTER_FACTOR
coef_matrix(i,:) = coef_matrix(CX,:);
end
end
%check if input is integer
int_sti=round(stimulation);
T = (int_sti ~= stimulation);
if (max(T)~=0)
fprintf("WARNING : Integer Input Expected: Rounding Fractional Input to Nearest Integer... ");
end
%Input overflow check
switch data_type
case 1
%set max/min for signed
maxdat = 2^(DATA_WIDTH-1)-1;
mindat = -maxdat-1;
case 2
%set max/min for unsigned
maxdat = 2^DATA_WIDTH-1;
mindat = 0;
end
if(data_type == 2)
if(abs(coef_matrix) == coef_matrix)
FIR_WIDTH = FIR_WIDTH +1;
end
end
%Saturating Input Value
a=find(int_sti>maxdat);
b=find(int_sti<mindat);
if (~isempty(a)|~isempty(b))
fprintf("WARNING : Input Amplitude Exceeds MAXIMUM/MINIMUM allowable values - saturating input values... ");
lena = length (a);
lenb = length (b);
for i =1:lena
fprintf("%d > %d ", int_sti(a(i)), maxdat);
int_sti(a(i)) = maxdat;
end
for i =1:lenb
fprintf("%d < %d ", int_sti(b(i)), mindat);
int_sti(b(i)) = mindat;
end
end
% Add interpolation
inter_sti = zeros(1, INTER_FACTOR * length(int_sti));
inter_sti(1:INTER_FACTOR:INTER_FACTOR * length(int_sti)) = int_sti;
for i = 1 : DY *INTER_FACTOR
coef_current = coef_matrix(i,:);
output_temp(i) = simp_adaptive (inter_sti, coef_current, i);
end
% Truncate output
len1 = length(output_temp);
switch LSB_TYPE
case 0
%truncate
out_dec = bi_trunc_lsb(output_temp,LSB_RM,FIR_WIDTH);
case 1
%round
out_dec = bi_round(output_temp,LSB_RM, FIR_WIDTH);
end
switch MSB_TYPE
case 0
%truncate
out_dec = bi_trunc_msb(out_dec,MSB_RM,FIR_WIDTH-LSB_RM);
case 1
%round
out_dec = bi_satu(out_dec,MSB_RM, FIR_WIDTH-LSB_RM);
end
% choose decimation output in phase=DECI_FACTOR-1
if(DECI_FACTOR == 1)
output = out_dec;
else
output = out_dec(DECI_FACTOR:DECI_FACTOR:len1);
end
function[output, outindex] = simp_adaptive (int_sti, coef_current, data_index, output)
%Simulation is the whole input sequence
%coef_current is the current coefficient set
%data_index gives the last data to use
%outputs are the sum of input and coef multiplication
%outindex is the next data_index
sti_current = zeros(length(coef_current),1);
data_length = length(int_sti);
%Check data index
if (data_index > data_length)
fprintf("ERROR: DATA INDEX IS LARGER THAN DATA LENGTH!!! ");
return;
end
for i = 1: length(coef_current)
if ((data_index -i+1)>0 & (data_index - i+1)<=data_length)
sti_current(i,1) = int_sti(data_index - i+1);
end
end
outindex= data_index+1;
output = coef_current * sti_current;
% end of function simp_adaptive
function output = bi_round(data_in,LSB_RM,ORI_WIDTH, output)
% LSB_RM is the bit to lose in LSB
% ORI_WIDTH is the original data width
data = round (data_in / 2^LSB_RM);
output = bi_satu(data,0,ORI_WIDTH - LSB_RM);
%end of function bi_trunc_lsb
function output = bi_trunc_lsb(data_in,LSB_RM,ORI_WIDTH, output)
% LSB_RM is the bit to lose in LSB
% ORI_WIDTH is the original data width
%2"s complement system
output = bitshift(2^ORI_WIDTH*(data_in<0) + data_in, -LSB_RM) - 2^(ORI_WIDTH-LSB_RM) *(data_in<0);
% end of function bi_round
function output = bi_trunc_msb(data_in,MSB_RM,ORI_WIDTH, output)
% MSB_RM is the bit to lose in LSB
% ORI_WIDTH is the original data width
%2"s complement system
data = 2^ORI_WIDTH * (data_in < 0)+ data_in;
erase_num = 2^(ORI_WIDTH - MSB_RM) - 1;
data = bitand(data, erase_num);
output = data - 2^(ORI_WIDTH - MSB_RM)*(bitget(data,ORI_WIDTH - MSB_RM));
%end of bi_trunc_msb
function output = bi_satu(data_in,MSB_RM,ORI_WIDTH, output)
% MSB_RM is the bit to lose in LSB
% ORI_WIDTH is the original data width
%2"s complement system
maxdat = 2^(ORI_WIDTH - MSB_RM -1)-1;
mindat = 2^(ORI_WIDTH - MSB_RM -1)*(-1);
data_in(find(data_in > maxdat)) = maxdat;
data_in(find(data_in < mindat)) = mindat;
output = data_in;
%end of bi_satu
- 左迁
-
fdatool里可以选择参数设计,并输出.m文件或VHDL