pi/4dqpsk的matlab及FPGA仿真

最新推荐文章于 2024-09-28 23:01:02 发布

Adolfor

最新推荐文章于 2024-09-28 23:01:02 发布

阅读量6.1k

点赞数 11

分类专栏： 4dqpsk 文章标签：信号处理 verilog matlab

本文链接：https://blog.csdn.net/Adolfor/article/details/106892071

版权

4dqpsk 专栏收录该内容

1 篇文章 2 订阅

订阅专栏

基于八相偏移调制的原理，本着资料稀少的原则，分享一下自己的经验。

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% PIQPSKMODEM使用清单
% 卷积编码/编码的Ser/Ber~Snr曲线
% 仿真星座图，pi/4_QPSK信号的产生方法基带及调制；
% 仿真差分解调 QPSK信号的过程及结果；
% 符号速率 ＝1Mbps；
% 基带成形滤波器滚降系数a=0.5；
% 采样速率 ＝8Mbps； 
% 绘制Hilbert滤波器及普通带通滤波器的频率响应图
% 绘制 pi/4_QPSK信号的频谱及时域波形，绘制解调后的同相正交支路眼图
% 暂缺同步
close all;
clear all;
clc;
ps=1*10^6;   %码速率为1MHz
a=0.8;       %成形滤波器系数
B=(1+a)*ps;  %中频信号处理带宽
Fs=8*10^6;   %采样速率
fc=2*10^6;   %载波频率
N=2000;     %仿真数据的长度

t=0:1/Fs:(N*Fs/ps-1)/Fs;%产生长度为N,频率为fs的时间序列
s=randi(4,1,N);       %产生随机四进制数据作为原始数据
A=[1i,2^0.5/2+2^0.5/2*1i,1,2^0.5/2-2^0.5/2*1i,-1i,-2^0.5/2-2^0.5/2*1i,-1,-2^0.5/2+2^0.5/2*1i];
Maptable=[0 0;0 1;1 1;1 0];
figure ;
scatter(real(A),imag(A),'filled');title('pi/4QPSK星座图');

%将绝对码转化为二进制绝对码
bin_s=zeros(1,2*N);
for i=1:1:N
  if(s(i)==1)bin_s(2*i-1)=0,bin_s(2*i)=0;
  else if(s(i)==2)bin_s(2*i-1)=0,bin_s(2*i)=1;
      else if(s(i)==3)bin_s(2*i-1)=1,bin_s(2*i)=1;
          else if(s(i)==4)bin_s(2*i-1)=1,bin_s(2*i)=0;
            end
          end
      end
  end
end
%对二进制绝对码进行卷积
Tre = poly2trellis(7,[133 171]);
code_bin_s= convenc(bin_s,Tre);

%将卷积码转换为四进制
code_s=zeros(1,2*N);
for i=1:2*N
      if(code_bin_s(2*i-1)==0&&code_bin_s(2*i)==0)code_s(i)=1;end
      if(code_bin_s(2*i-1)==0&&code_bin_s(2*i)==1)code_s(i)=2;end
      if(code_bin_s(2*i-1)==1&&code_bin_s(2*i)==1)code_s(i)=3;end
      if(code_bin_s(2*i-1)==1&&code_bin_s(2*i)==0)code_s(i)=4;end
end
%将绝对码变换为相对码
xk=ones(1,N);
yk=ones(1,N);

for i=2:N
    if s(i)==1
        xk(i)=xk(i-1)*cos(pi/4)-yk(i-1)*sin(pi/4);
        yk(i)=yk(i-1)*cos(pi/4)+xk(i-1)*sin(pi/4);
    elseif s(i)==2 
        xk(i)=xk(i-1)*cos(-pi/4)-yk(i-1)*sin(-pi/4);
        yk(i)=yk(i-1)*cos(-pi/4)+xk(i-1)*sin(-pi/4);
    elseif s(i)==3 
        xk(i)=xk(i-1)*cos(3*pi/4)-yk(i-1)*sin(3*pi/4);
        yk(i)=yk(i-1)*cos(3*pi/4)+xk(i-1)*sin(3*pi/4);
    elseif s(i)==4 
        xk(i)=xk(i-1)*cos(-3*pi/4)-yk(i-1)*sin(-3*pi/4);
        yk(i)=yk(i-1)*cos(-3*pi/4)+xk(i-1)*sin(-3*pi/4);
    end
end
%将卷积绝对码转化为卷积相对码
xkk=ones(1,2*N);
ykk=ones(1,2*N);
for i=2:2*N
    if code_s(i)==1
        xkk(i)=xkk(i-1)*cos(pi/4)-ykk(i-1)*sin(pi/4);
        ykk(i)=ykk(i-1)*cos(pi/4)+xkk(i-1)*sin(pi/4);
    elseif code_s(i)==2
        xkk(i)=xkk(i-1)*cos(-pi/4)-ykk(i-1)*sin(-pi/4);
        ykk(i)=ykk(i-1)*cos(-pi/4)+xkk(i-1)*sin(-pi/4);
    elseif code_s(i)==3 
        xkk(i)=xkk(i-1)*cos(3*pi/4)-ykk(i-1)*sin(3*pi/4);
        ykk(i)=ykk(i-1)*cos(3*pi/4)+xkk(i-1)*sin(3*pi/4);
    elseif code_s(i)==4
        xkk(i)=xkk(i-1)*cos(-3*pi/4)-ykk(i-1)*sin(-3*pi/4);
        ykk(i)=ykk(i-1)*cos(-3*pi/4)+xkk(i-1)*sin(-3*pi/4);
    end
end
%对相对码数据以Fs频率采样
Ads_i=upsample(xk,Fs/ps);
Ads_q=upsample(yk,Fs/ps);

%加噪声
SNR=20;
Ads_i=awgn(Ads_i,SNR);
Ads_q=awgn(Ads_q,SNR);

%设计平方根升余弦滤波器
n_T=[-2 2];
rate=Fs/ps;
T=1;
Shape_b = rcosfir(a,n_T,rate,T,'sqrt');
%对采样后的数据进行升余弦滤波;
rcos_Ads_i=filter(Shape_b,1,Ads_i);
rcos_Ads_q=filter(Shape_b,1,Ads_q);

%产生同相正交两路载频信号
f0_i=cos(2*pi*fc*t); 
f0_q=sin(2*pi*fc*t);       

%产生PI/4_QPSK已调信号
piqpsk=rcos_Ads_i.*f0_i-rcos_Ads_q.*f0_q;       

dd=(xk+yk*1i)*(2^0.5)/2;
Ps=1;
Pn=Ps/SNR;
z=sqrt(Pn/2)*(randn(size(dd))+1i*randn(size(dd)));
y=dd+z;
figure;scatter(real(y(1:N/2)),imag(y(1:N/2)),'filled');title('π/4QPSK接收星座图');

%设计Hilbert滤波器及相同阶数的普通带通滤波器
fpm=[0 0.25 1 3 3.75 4]*10^6*2/Fs;  %firpm函数的频段向量
magpm=[0 0 1 1 0 0];                %firpm函数的幅值向量
n=30;                               %滤波器阶数
h_bpf=firpm(n,fpm,magpm,'hilbert滤波器') ;%Hilbert带通滤波器
bpf=firpm(n,fpm,magpm);             %普通带通滤波器

figure;
%绘制Hilbert滤波器及普通带通滤波器频率响应
freqz(h_bpf);
freqz(bpf);

%完成对PI/4_QPSK信号的Hilbert滤波及普通滤波
piqpsk_i=filter(bpf,1,piqpsk);
piqpsk_q=filter(h_bpf,1,piqpsk);

%对普通带通滤波后的数据进行一个符号周期延时处理
piqpsk_di=[zeros(1,Fs/ps),piqpsk_i(1:length(piqpsk_i)-Fs/ps)];

%实现差分解调
demod_mult_i=piqpsk_i.*piqpsk_di;
demod_mult_q=piqpsk_q.*piqpsk_di;

%对乘法运算后的同相正交支路滤波
demod_i=filter(Shape_b,1,demod_mult_i);
demod_q=filter(Shape_b,1,demod_mult_q);

%绘制解调后的同相正交支路眼图
eyediagram(demod_i,4*Fs/ps)
eyediagram(demod_q,4*Fs/ps)
%得到的解码信号demod_i=sing,g为相位，demod_q为cosg

%绘制pi4_QPSK信号频谱、pi4_QPSK信号时域波形
figure;
m_piqpsk=20*log10(abs(fft(piqpsk,1024)));m_piqpsk=m_piqpsk-max(m_piqpsk);
%设置幅频响应的横坐标单位为MHz
x_f=[0:(Fs/length(m_piqpsk)):Fs/2];x_f=x_f/10^6;
%只显示正频率部分的幅频响应
mpiqpsk=m_piqpsk(1:length(x_f));
%设置时域波表的横坐标单位为us

Len=100;%设置时域波形显示的点数
x_t=1:Len;%产生长度为Len的时间序列
x_t=x_t/Fs*10^6;
%显示所需的频谱及时域波形
subplot(211); plot(x_f,mpiqpsk);       
legend('PI/4 QPSK信号频谱');
xlabel('频率(MHz)');ylabel('幅度(dB)');grid on;
subplot(212);plot(x_t,piqpsk(101:Len+100));
legend('PI/4 QPSK时域信号波形');
xlabel('时间(us)');ylabel('幅度(V)');grid on;

% figure;
% subplot(211);
% x_t=1:8:1000;
% plot(x_t,demod_i(x_t));title("解调后同相基带波形");
% subplot(212);
% plot(x_t,demod_q(x_t));title("解调后正交基带波形");
% % 

%进行信号解调，统计不同信噪比下的误码率与误符号率，并于理论结果对比验证
sum=2;
H=cos(pi/4);
for i=2:length(dd)
     B=[ real(dd(i-1))*H-imag(dd(i-1))*H+(imag(dd(i-1))*H+real(dd(i-1))*H)*1i,
         real(dd(i-1))*H-imag(dd(i-1))*(-H)+(imag(dd(i-1))*H+real(dd(i-1))*(-H))*1i,
         real(dd(i-1))*(-H)-imag(dd(i-1))*H+(imag(dd(i-1))*(-H)+real(dd(i-1))*H)*1i,
         real(dd(i-1))*(-H)-imag(dd(i-1))*(-H)+(imag(dd(i-1))*(-H)+real(dd(i-1))*(-H))*1i ];
     Dist=abs(dd(i)-B);      
    [~,Index]=min(Dist);      %找到距离最近的星座图点 
    x_est(i)=B(Index);        %得到发射调制信号的估计
    result(sum)=Index;
    sum=sum+1;
    bin_res(2*i-1:2*i)=Maptable(Index,:); %得到比特数据的估计
end

figure;
Len=N;
x_t=1:Len;
plot(x_t,bin_res(x_t));
title("PI/4QPSK解调后信号");


Ber1=ones(1,20);Ber2=ones(1,20);
num=0;
for snr=1:20;                 %信噪比循环,实际使用请改成-10到10 ，SNR_DB改为snr-21
    for mt=1:100              %蒙特卡洛循环,因省时实际使用时设置为1000，num也除以1000
        SNR_dB=snr;        %不同信噪比下生成AWGN信道 
        SNR=10^(SNR_dB/10);   
        Ps=1;  
        Pn=Ps/SNR;
        z=sqrt(Pn/2)*(randn(size(dd))+1i*randn(size(dd)));
        y=dd+z;
        for i=2:1:length(y)
     B=[ real(y(i-1))*H-imag(y(i-1))*H+(imag(y(i-1))*H+real(y(i-1))*H)*1i,
         real(y(i-1))*H-imag(y(i-1))*(-H)+(imag(y(i-1))*H+real(y(i-1))*(-H))*1i,
         real(y(i-1))*(-H)-imag(y(i-1))*H+(imag(y(i-1))*(-H)+real(y(i-1))*H)*1i,
         real(y(i-1))*(-H)-imag(y(i-1))*(-H)+(imag(y(i-1))*(-H)+real(y(i-1))*(-H))*1i ];
     Dist=abs(y(i)-B);      
    [~,Index]=min(Dist);      %找到距离最近的星座图点 
    result(i)=Index;
    bin_res(2*i-1:2*i)=Maptable(Index,:); %得到比特数据的估计
        end
        [tureerr1,per2]=symerr(result,s);  %per2单次过程中的误符号率
        num=num+per2;          %求100次循环误码率总和
    end
 Ber1(snr)=num/100;
 num=0;
   %100次误码率的平均值
    %theoerr1(snr)=2*qfunc(sqrt(2*SNR)*sin(pi/4));
    theoerr1(snr)=1/2*erfc(sqrt(SNR/2));
end
Ser1=Ber1*log2(2);

%%卷积后的Ber/Ser~SNR曲线

num=0;sum=1;
code_dd=xkk+1i*ykk;
for snr=1:20                  %信噪比循环
    for mt=1:100              %蒙特卡洛循环，实际使用时设置为1000，num也除以1000
        SNR_dB=snr;        %不同信噪比下生成AWGN信道 
        SNR=10^(SNR_dB/10);   
        Ps=1;  
        Pn=Ps/SNR;
        z=sqrt(Pn/2)*(randn(size(code_dd))+1i*randn(size(code_dd)));
        y=code_dd+z;
        for i=2:1:length(y)
     B=[ real(y(i-1))*H-imag(y(i-1))*H+(imag(y(i-1))*H+real(y(i-1))*H)*1i,
         real(y(i-1))*H-imag(y(i-1))*(-H)+(imag(y(i-1))*H+real(y(i-1))*(-H))*1i,
         real(y(i-1))*(-H)-imag(y(i-1))*H+(imag(y(i-1))*(-H)+real(y(i-1))*H)*1i,
         real(y(i-1))*(-H)-imag(y(i-1))*(-H)+(imag(y(i-1))*(-H)+real(y(i-1))*(-H))*1i ];
     Dist=abs(y(i)-B);      
    [~,Index]=min(Dist);   
    code_bin_res(2*i-1:2*i)=Maptable(Index,:); %得到二进制解码信号
        end
        tblen = 18;
        code_res = vitdec(code_bin_res,Tre,tblen,'trunc','hard');    
        [tureerr2,per1]=symerr(code_res,bin_s);  %per2单次过程中的误符号率
        num=num+per1;          %求100次循环误码率总和
    end
    Ber2(snr)=num/100; 
       num=0;
    theoerr1(snr)=2*qfunc(sqrt(2*SNR)*sin(pi/4));
    theoerr2(snr)=1/2*erfc(sqrt(SNR/2));
end

 Ser2=Ber2*log2(2);
S=1:20; %信噪比区间
figure;subplot(211);
semilogy(S,Ber1,'r',S,theoerr1*2,'b');
xlabel('SNR');ylabel('Ber');title('编码Ber/SNR的关系');legend('实际','理论');grid on;  
subplot(212);
semilogy(S,Ber2,'g',S,theoerr2*2,'b');
xlabel('SNR');ylabel('Ber');title('卷积编码Ber/SNR的关系');
legend('实际','理论');grid on;  

figure;subplot(211);
semilogy(S,Ser1,'r',S,theoerr1,'b');
xlabel('SNR');ylabel('Ser');title('编码Ser/SNR的关系');legend('实际','理论');grid on;
subplot(212);
semilogy(S,Ser2,'g',S,theoerr2,'b');
xlabel('SNR');ylabel('Ser');title('卷积编码Ser/SNR的关系');
legend('实际','理论');grid on;

采用中频差分解调，下面为基带差分解调代码demo

close all;
clc;
ps=4*10^6;   %码速率
a=0.8;       %成形滤波器系数
B=(1+a)*ps;  %中频信号处理带宽
Fs=64*10^6;   %采样速率
fc=2.4*10^9;   %载波频率
N=2000;     %仿真数据的长度

t=0:1/Fs:(N*Fs/ps-1)/Fs;%产生长度为N,频率为fs的时间序列
s=(randi(4,1,N)-1);       %产生随机四进制数据作为原始数据
A=[1i,2^0.5/2+2^0.5/2*1i,1,2^0.5/2-2^0.5/2*1i,-1i,-2^0.5/2-2^0.5/2*1i,-1,-2^0.5/2+2^0.5/2*1i];
Maptable=[0 0;0 1;1 1;1 0];
% figure ;
% scatter(real(A),imag(A),'filled');title('pi/4QPSK星座图');

%将绝对码变换为相对码
xk=ones(1,N);
yk=ones(1,N);

for i=2:N
    if s(i)==3
        xk(i)=xk(i-1)*cos(pi/4)-yk(i-1)*sin(pi/4);
        yk(i)=yk(i-1)*cos(pi/4)+xk(i-1)*sin(pi/4);
    elseif s(i)==2 
        xk(i)=xk(i-1)*cos(-pi/4)-yk(i-1)*sin(-pi/4);
        yk(i)=yk(i-1)*cos(-pi/4)+xk(i-1)*sin(-pi/4);
    elseif s(i)==1 
        xk(i)=xk(i-1)*cos(3*pi/4)-yk(i-1)*sin(3*pi/4);
        yk(i)=yk(i-1)*cos(3*pi/4)+xk(i-1)*sin(3*pi/4);
    elseif s(i)==0 
        xk(i)=xk(i-1)*cos(-3*pi/4)-yk(i-1)*sin(-3*pi/4);
        yk(i)=yk(i-1)*cos(-3*pi/4)+xk(i-1)*sin(-3*pi/4);
    end
end

%对相对码数据以Fs频率采样
Ads_i=upsample(xk,Fs/ps);
Ads_q=upsample(yk,Fs/ps);

%加噪声
SNR=50;
Ads_i=awgn(Ads_i,SNR);
Ads_q=awgn(Ads_q,SNR);

%设计平方根升余弦滤波器
n_T=[-2 2];
rate=Fs/ps;
T=1;
Shape_b = rcosfir(a,n_T,rate,T,'sqrt');
%对采样后的数据进行升余弦滤波;
rcos_Ads_i=filter(Shape_b,1,Ads_i);
rcos_Ads_q=filter(Shape_b,1,Ads_q);

%-------------匹配滤波+降采样------------
rcos_Ads_i=upfirdn(rcos_Ads_i,Shape_b,1,16);
rcos_Ads_q=upfirdn(rcos_Ads_q,Shape_b,1,16);

[j,k]=size(rcos_Ads_i);
%不滤波直接采样
w=zeros(1,k);
z=zeros(1,k);
i1=zeros(1,k);
q1=zeros(1,k);
so=zeros(1,k);

w=rcos_Ads_i;
z=rcos_Ads_q;
for i=2:1:k
    i0(i)=w(i)*w(i-1)+z(i)*z(i-1);
    q0(i)=z(i)*w(i-1)-w(i)*z(i-1);
    if i0(i)>0&&q0(i)>0
        so(i)=3;
    elseif i0(i)<0&&q0(i)>0
        so(i)=1;
    elseif i0(i)<0&&q0(i)<0
        so(i)=0;
    elseif i0(i)>0&&q0(i)<0
        so(i)=2;
    end 
end
%----------输出-----------
soo(1:N)=so(k-N+1:k);
%----------误码-----------
a=0;
for i=1:2000
    if s(i)==soo(i)
        a=a+1;
    end
end

采用FGPA仿真则分为不同模块，其中本人忽略了3个。
1.串并转换
2.差分编码
3.载波调制（忽略）
4.接收载波（忽略）
5.恢复相位（忽略）
6.延迟相乘
7.门限判决（位同步）
8.并串转换
IP核：fir滤波器（cosfir）移位寄存器以及有符号乘法器
其中给出差分编码代码如下（也可八位量化）VerilogHDL

module PiQpskCode ( 
	rst,clk,din,
	xk,yk); 
	
	input		rst;   
	input		clk;   
	input	wire [1:0]	din;    
	
	output reg [9:0]xk;
	output reg [9:0]yk;
	reg[1:0]cuu=2'b0;
	reg [2:0]addr=3'b0;
	reg [2:0]delt_phase;
	reg cclk=0;
	parameter A = 10'b0110100100;//??0,1???ò????·????ù??????·ù?????????±???ó????????420±?,???ó????????????
	parameter B = 10'b0100101001;
	parameter n = -10'b0000000001;
	parameter ZERO = 10'b0000000000;
	
always@(posedge clk)begin
cuu=cuu+2'b1;
if(cuu==2'b01)cclk=~cclk;
end
	always@(posedge cclk )
				case(din)
					2'b00:	delt_phase = 3'b101;
					2'b01:	delt_phase = 3'b011;
					2'b10:	delt_phase = 3'b001;
					default:delt_phase = 3'b111;
				endcase
	
			

/*
	assign delt_phase = 	(din == 2'b00)?3'b101:
								(din == 2'b01)?3'b011:
								(din == 2'b10)?3'b001: 3'b111;
								
	always@(posedge clk or posedge rst)
		if(rst)
			addr = 3'b0;
		else 
			addr = addr + delt_phase;
	*/
	always@(posedge cclk or posedge rst)
		begin
			if(rst)
				begin
				
					xk<=10'b0000000000;
					yk<=10'b0000000000;
				end
			else begin addr = addr + delt_phase;
					case(addr)
						3'b000:
							begin
								xk <= A;
								yk <=ZERO;
							end
						3'b001:
							begin
								xk <= B;
								yk <=B;
							end
						3'b010:
							begin
								xk <= ZERO;
								yk <=A;
							end
						3'b011:
							begin
								xk <= n*B;
								yk <=B;
							end
						3'b100:
							begin
								xk <= n*A;
								yk <=ZERO;
							end
						3'b101:
							begin
								xk <= n*B;
								yk <=n*B;
							end
						3'b110:
							begin
								xk <= ZERO;
								yk <=n*A;
							end
						3'b111:
							begin
								xk <= B;
								yk <= n*B;
							end
						default :
							begin
								xk <= 10'bxxxxxxxxxx;
								yk <= 10'bxxxxxxxxxx;
							end
					endcase
				end
				end
				endmodule

其中给出main代码如下

`timescale 1ns / 1ns
//
// Company: 
// Engineer: 
// 
// Create Date: 2020/05/19 15:00:48
// Design Name: 
// Module Name: main
// Project Name: 
// Target Devices: 
// Tool Versions: 
// Description: 
// 
// Dependencies: 
// 
// Revision:
// Revision 0.01 - File Created
// Additional Comments:
// 
//
module main(rst,clk,aclk,clk4,din,di_out,dq_out,di_lpf,dq_lpf);
    input		rst;  
	input		clk;  
	input     clk4;
	input     aclk; 
	input	   din;  
	output  signed[31:0]di_lpf;
    output  signed[31:0]dq_lpf;
    output  signed[9:0]di_out;
    output  signed[9:0]dq_out;
 
    wire [9:0]	di; 
    wire [9:0]	dq; 
   
    //串/并转换，2 Mbps的单比特数据转换为1 Mbps的双比特码元
	 reg [1:0] ab;
	 reg [2:0] count;
	 reg dint;
    always @(posedge clk or posedge rst)
	    if (rst)
		    begin
			    ab <= 2'd0;
				 count <= 3'd0;
				 dint <= 1'b0;
			end
		else
		   begin  
		      count <= count + 3'd1;
			   if (count ==3'd0)
					dint <= din;
				else if (count == 3'd4)
					ab <= {din,dint};
			end
          //基带编码
		PiQpskCode u1 (
	   .rst (rst),
		.clk (clk),
		.din (ab),
		.xk (di),
		.yk (dq));

	wire source_validi,source_validq;
	wire signed [31:0] di_lpf;	
	wire signed [31:0] dq_lpf;	
	wire axis_tready,axis_treadyy;
	
   fir_compiler_0 u2(
		.aclk (aclk),
		.s_axis_data_tdata (di),
		.s_axis_data_tvalid (1'b1),
		.s_axis_data_tready (axis_tready),
		.m_axis_data_tdata(di_lpf),
		.m_axis_data_tvalid (source_validi));
		
	fir_compiler_0 u3(
		.aclk (aclk),
		.s_axis_data_tdata (dq),
		.s_axis_data_tvalid (1'b1),
		.s_axis_data_tready (axis_treadyy),
		.m_axis_data_tdata(dq_lpf),
		.m_axis_data_tvalid (source_validq));
	
	wire dout;
	wire [1:0]bitsync;	
	jietiao (
	.rst(rst),.clk(clk),.clk4(clk4),.di(di_lpf),.dq(dq_lpf),
	.dout(dout),.bitsync(bitsync)); 
	
endmodule

其中位同步代码分为几个模块可以加也可不加，如果想要的话可以直接下载我的资源。代码除参考全部由本人撰写，参考<数字调制解调技术的MATLAB与FPGA实现：ALTERA／VERILOG>因为最近quartus被收购所以换了vivado，涉及到ip核的设计，如果不会弄全部代码依旧可以去下载我的资源，谢谢。

lpf.coe:

radix = 10;
coefdata=
          27
        22
        16
         8
        -0
        -9
       -19
       -28
       -37
       -44
       -51
       -55
       -57
       -56
       -52
       -46
       -37
       -25
       -11
         4
        21
        37
        54
        69
        82
        93
       100
       103
       101
        94
        83
        66
        45
        20
        -8
       -38
       -70
      -101
      -131
      -157
      -179
      -195
      -204
      -204
      -194
      -174
      -144
      -103
       -51
        11
        81
       160
       245
       334
       426
       518
       609
       695
       775
       847
       908
       958
       994
      1017
      1024
      1017
       994
       958
       908
       847
       775
       695
       609
       518
       426
       334
       245
       160
        81
        11
       -51
      -103
      -144
      -174
      -194
      -204
      -204
      -195
      -179
      -157
      -131
      -101
       -70
       -38
        -8
        20
        45
        66
        83
        94
       101
       103
       100
        93
        82
        69
        54
        37
        21
         4
       -11
       -25
       -37
       -46
       -52
       -56
       -57
       -55
       -51
       -44
       -37
       -28
       -19
        -9
        -0
         8
        16
        22
        27


;

fdatool制作的余弦滚降滤波器。
位同步技术