FFT算法的完整DSP实现
傅里叶变换或者FFT的理论参考:
[1] http://www.dspguide.com/ch12/2.htm
The Scientist and Engineer's Guide to Digital Signal Processing, By Steven W. Smith, Ph.D.
[2] http://blog.csdn.net/v_JULY_v/article/details/6196862,可当作[1]的中文参考
[3] 任意一本数字信号处理教材,上面都有详细的推导DCT求解转换为FFT求解的过程
[4] TI文档:基于TMS320C64x+DSP的FFT实现。 使用baidu/google可以搜索到。
另外,FFT的开源代码可参考:
[1] FFTW: http://www.fftw.org/ 最快,最好的开源FFT。
[2] FFTReal: http://ldesoras.free.fr/prod.html#src_fftreal 轻量级FFT算法实现
[3] KISS FFT: http://sourceforge.net/projects/kissfft/ 简单易用的FFT的C语言实现
1. 有关FFT理论的一点小小解释
关于FFT这里只想提到两点:
(1)DFT变换对的表达式(必须记住)
—— 称旋转因子
(2)FFT用途——目标只有一个,加速DFT的计算效率。
DFT计算X(k)需要N^2次复数乘法和N(N-1)次复数加法;FFT将N^2的计算量降为。
“FFT其实是很难的东西,即使常年在这个领域下打拼的科学家也未必能很好的写出FFT的算法。”——摘自参考上面提供的参考文献[1]
因此,我们不必太过纠结于细节,当明白FFT理论后,将已有的算法挪过来用就OK了,不必为闭着教材写不出FFT而郁闷不堪。
FFT的BASIC程序伪代码如下:
IFFT的BASIC程序伪代码如下(IFFT通过调用FFT计算):
FFT算法的流程图如下图,总结为3过程3循环:
(1)3过程:单点时域分解(倒位序过程) + 单点时域计算单点频谱 + 频域合成
(2)3循环:外循环——分解次数,中循环——sub-DFT运算,内循环——2点蝶形算法
分解过程或者说倒位序的获得参考下图理解:
2. FFT的DSP实现
下面为本人使用C语言实现的FFT及IFFT算法实例,能计算任意以2为对数底的采样点数的FFT,算法参考上面给的流程图。
- /*
- * zx_fft.h
- *
- * Created on: 2013-8-5
- * Author: monkeyzx
- */
- #ifndef ZX_FFT_H_
- #define ZX_FFT_H_
- typedef float FFT_TYPE;
- #ifndef PI
- #define PI (3.14159265f)
- #endif
- typedef struct complex_st {
- FFT_TYPE real;
- FFT_TYPE img;
- } complex;
- int fft(complex *x, int N);
- int ifft(complex *x, int N);
- void zx_fft(void);
- #endif /* ZX_FFT_H_ */
/** zx_fft.h** Created on: 2013-8-5* Author: monkeyzx*/#ifndef ZX_FFT_H_
#define ZX_FFT_H_typedef float FFT_TYPE;#ifndef PI
#define PI (3.14159265f)
#endiftypedef struct complex_st {FFT_TYPE real;FFT_TYPE img;
} complex;int fft(complex *x, int N);
int ifft(complex *x, int N);
void zx_fft(void);#endif /* ZX_FFT_H_ */
- /*
- * zx_fft.c
- *
- * Implementation of Fast Fourier Transform(FFT)
- * and reversal Fast Fourier Transform(IFFT)
- *
- * Created on: 2013-8-5
- * Author: monkeyzx
- */
- #include "zx_fft.h"
- #include <math.h>
- #include <stdlib.h>
- /*
- * Bit Reverse
- * === Input ===
- * x : complex numbers
- * n : nodes of FFT. @N should be power of 2, that is 2^(*)
- * l : count by bit of binary format, @l=CEIL{log2(n)}
- * === Output ===
- * r : results after reversed.
- * Note: I use a local variable @temp that result @r can be set
- * to @x and won't overlap.
- */
- static void BitReverse(complex *x, complex *r, int n, int l)
- {
- int i = 0;
- int j = 0;
- short stk = 0;
- static complex *temp = 0;
- temp = (complex *)malloc(sizeof(complex) * n);
- if (!temp) {
- return;
- }
- for(i=0; i<n; i++) {
- stk = 0;
- j = 0;
- do {
- stk |= (i>>(j++)) & 0x01;
- if(j<l)
- {
- stk <<= 1;
- }
- }while(j<l);
- if(stk < n) { /* 满足倒位序输出 */
- temp[stk] = x[i];
- }
- }
- /* copy @temp to @r */
- for (i=0; i<n; i++) {
- r[i] = temp[i];
- }
- free(temp);
- }
- /*
- * FFT Algorithm
- * === Inputs ===
- * x : complex numbers
- * N : nodes of FFT. @N should be power of 2, that is 2^(*)
- * === Output ===
- * the @x contains the result of FFT algorithm, so the original data
- * in @x is destroyed, please store them before using FFT.
- */
- int fft(complex *x, int N)
- {
- int i,j,l,ip;
- static int M = 0;
- static int le,le2;
- static FFT_TYPE sR,sI,tR,tI,uR,uI;
- M = (int)(log(N) / log(2));
- /*
- * bit reversal sorting
- */
- BitReverse(x,x,N,M);
- /*
- * For Loops
- */
- for (l=1; l<=M; l++) { /* loop for ceil{log2(N)} */
- le = (int)pow(2,l);
- le2 = (int)(le / 2);
- uR = 1;
- uI = 0;
- sR = cos(PI / le2);
- sI = -sin(PI / le2);
- for (j=1; j<=le2; j++) { /* loop for each sub DFT */
- //jm1 = j - 1;
- for (i=j-1; i<=N-1; i+=le) { /* loop for each butterfly */
- ip = i + le2;
- tR = x[ip].real * uR - x[ip].img * uI;
- tI = x[ip].real * uI + x[ip].img * uR;
- x[ip].real = x[i].real - tR;
- x[ip].img = x[i].img - tI;
- x[i].real += tR;
- x[i].img += tI;
- } /* Next i */
- tR = uR;
- uR = tR * sR - uI * sI;
- uI = tR * sI + uI *sR;
- } /* Next j */
- } /* Next l */
- return 0;
- }
- /*
- * Inverse FFT Algorithm
- * === Inputs ===
- * x : complex numbers
- * N : nodes of FFT. @N should be power of 2, that is 2^(*)
- * === Output ===
- * the @x contains the result of FFT algorithm, so the original data
- * in @x is destroyed, please store them before using FFT.
- */
- int ifft(complex *x, int N)
- {
- int k = 0;
- for (k=0; k<=N-1; k++) {
- x[k].img = -x[k].img;
- }
- fft(x, N); /* using FFT */
- for (k=0; k<=N-1; k++) {
- x[k].real = x[k].real / N;
- x[k].img = -x[k].img / N;
- }
- return 0;
- }
- /*
- * Code below is an example of using FFT and IFFT.
- */
- #define SAMPLE_NODES (128)
- complex x[SAMPLE_NODES];
- int INPUT[SAMPLE_NODES];
- int OUTPUT[SAMPLE_NODES];
- static void MakeInput()
- {
- int i;
- for ( i=0;i<SAMPLE_NODES;i++ )
- {
- x[i].real = sin(PI*2*i/SAMPLE_NODES);
- x[i].img = 0.0f;
- INPUT[i]=sin(PI*2*i/SAMPLE_NODES)*1024;
- }
- }
- static void MakeOutput()
- {
- int i;
- for ( i=0;i<SAMPLE_NODES;i++ )
- {
- OUTPUT[i] = sqrt(x[i].real*x[i].real + x[i].img*x[i].img)*1024;
- }
- }
- void zx_fft(void)
- {
- MakeInput();
- fft(x,128);
- MakeOutput();
- ifft(x,128);
- MakeOutput();
- }
/** zx_fft.c** Implementation of Fast Fourier Transform(FFT)* and reversal Fast Fourier Transform(IFFT)** Created on: 2013-8-5* Author: monkeyzx*/#include "zx_fft.h"
#include <math.h>
#include <stdlib.h>/** Bit Reverse* === Input ===* x : complex numbers* n : nodes of FFT. @N should be power of 2, that is 2^(*)* l : count by bit of binary format, @l=CEIL{log2(n)}* === Output ===* r : results after reversed.* Note: I use a local variable @temp that result @r can be set* to @x and won't overlap.*/
static void BitReverse(complex *x, complex *r, int n, int l)
{int i = 0;int j = 0;short stk = 0;static complex *temp = 0;temp = (complex *)malloc(sizeof(complex) * n);if (!temp) {return;}for(i=0; i<n; i++) {stk = 0;j = 0;do {stk |= (i>>(j++)) & 0x01;if(j<l){stk <<= 1;}}while(j<l);if(stk < n) { /* 满足倒位序输出 */temp[stk] = x[i];}}/* copy @temp to @r */for (i=0; i<n; i++) {r[i] = temp[i];}free(temp);
}/** FFT Algorithm* === Inputs ===* x : complex numbers* N : nodes of FFT. @N should be power of 2, that is 2^(*)* === Output ===* the @x contains the result of FFT algorithm, so the original data* in @x is destroyed, please store them before using FFT.*/
int fft(complex *x, int N)
{int i,j,l,ip;static int M = 0;static int le,le2;static FFT_TYPE sR,sI,tR,tI,uR,uI;M = (int)(log(N) / log(2));/** bit reversal sorting*/BitReverse(x,x,N,M);/** For Loops*/for (l=1; l<=M; l++) { /* loop for ceil{log2(N)} */le = (int)pow(2,l);le2 = (int)(le / 2);uR = 1;uI = 0;sR = cos(PI / le2);sI = -sin(PI / le2);for (j=1; j<=le2; j++) { /* loop for each sub DFT *///jm1 = j - 1;for (i=j-1; i<=N-1; i+=le) { /* loop for each butterfly */ip = i + le2;tR = x[ip].real * uR - x[ip].img * uI;tI = x[ip].real * uI + x[ip].img * uR;x[ip].real = x[i].real - tR;x[ip].img = x[i].img - tI;x[i].real += tR;x[i].img += tI;} /* Next i */tR = uR;uR = tR * sR - uI * sI;uI = tR * sI + uI *sR;} /* Next j */} /* Next l */return 0;
}/** Inverse FFT Algorithm* === Inputs ===* x : complex numbers* N : nodes of FFT. @N should be power of 2, that is 2^(*)* === Output ===* the @x contains the result of FFT algorithm, so the original data* in @x is destroyed, please store them before using FFT.*/
int ifft(complex *x, int N)
{int k = 0;for (k=0; k<=N-1; k++) {x[k].img = -x[k].img;}fft(x, N); /* using FFT */for (k=0; k<=N-1; k++) {x[k].real = x[k].real / N;x[k].img = -x[k].img / N;}return 0;
}/** Code below is an example of using FFT and IFFT.*/
#define SAMPLE_NODES (128)
complex x[SAMPLE_NODES];
int INPUT[SAMPLE_NODES];
int OUTPUT[SAMPLE_NODES];static void MakeInput()
{int i;for ( i=0;i<SAMPLE_NODES;i++ ){x[i].real = sin(PI*2*i/SAMPLE_NODES);x[i].img = 0.0f;INPUT[i]=sin(PI*2*i/SAMPLE_NODES)*1024;}
}static void MakeOutput()
{int i;for ( i=0;i<SAMPLE_NODES;i++ ){OUTPUT[i] = sqrt(x[i].real*x[i].real + x[i].img*x[i].img)*1024;}
}void zx_fft(void)
{MakeInput();fft(x,128);MakeOutput();ifft(x,128);MakeOutput();
}
程序在TMS320C6713上实验,主函数中调用zx_fft()函数即可。
FFT的采样点数为128,输入信号的实数域为正弦信号,虚数域为0,数据精度定义FFT_TYPE为float类型,MakeInput和MakeOutput函数分别用于产生输入数据INPUT和输出数据OUTPUT的函数,便于使用CCS 的Graph功能绘制波形图。这里调试时使用CCS v5中的Tools -> Graph功能得到下面的波形图(怎么用自己琢磨,不会的使用CCS 的Help)。
输入波形
输入信号的频域幅值表示
FFT运算结果
对FFT运算结果逆变换(IFFT)
如何检验运算结果是否正确呢?有几种方法:
(1)使用matlab验证,下面为相同情况的matlab图形验证代码
- SAMPLE_NODES = 128;
- i = 1:SAMPLE_NODES;
- x = sin(pi*2*i / SAMPLE_NODES);
- subplot(2,2,1); plot(x);title('Inputs');
- axis([0 128 -1 1]);
- y = fft(x, SAMPLE_NODES);
- subplot(2,2,2); plot(abs(y));title('FFT');
- axis([0 128 0 80]);
- z = ifft(y, SAMPLE_NODES);
- subplot(2,2,3); plot(abs(z));title('IFFT');
- axis([0 128 0 1]);
SAMPLE_NODES = 128;
i = 1:SAMPLE_NODES;
x = sin(pi*2*i / SAMPLE_NODES);subplot(2,2,1); plot(x);title('Inputs');
axis([0 128 -1 1]);y = fft(x, SAMPLE_NODES);
subplot(2,2,2); plot(abs(y));title('FFT');
axis([0 128 0 80]);z = ifft(y, SAMPLE_NODES);
subplot(2,2,3); plot(abs(z));title('IFFT');
axis([0 128 0 1]);
(2)使用IFFT验证:输入信号的FFT获得的信号再IFFT,则的到的信号与原信号相同
可能大家发现输入信号上面的最后IFFT的信号似乎不同,这是因为FFT和IFFT存在精度截断误差(也叫数据截断噪声,意思就是说,我们使用的float数据类型数据位数有限,没法完全保留原始信号的信息)。因此,IFFT之后是复数(数据截断噪声引入了虚数域,只不过值很小),所以在绘图时使用了计算幅值的方法,
C代码中:
- OUTPUT[i] = sqrt(x[i].real*x[i].real + x[i].img*x[i].img)*1024;
OUTPUT[i] = sqrt(x[i].real*x[i].real + x[i].img*x[i].img)*1024;
matlab代码中:
- subplot(2,2,3); plot(abs(z));title('IFFT');
subplot(2,2,3); plot(abs(z));title('IFFT');
所以IFFT的结果将sin函数的负y轴数据翻到了正y轴。另外,在CCS v5的图形中我们将显示信号的幅度放大了1024倍便于观察,而matlab中没有放大。
=================
更正 更正 。。。
=================
上面程序中的BitReverse函数由于使用了malloc函数,当要分配的n比较大时,在DSP上运行会出现一定的问题,因此改用伪代码中提供的倒位序方法更可靠。
修正后的完整FFT代码文件粘贴如下,在实际的DSP项目中测试通过,可直接拷贝复用。
- /*
- * zx_fft.h
- *
- * Created on: 2013-8-5
- * Author: monkeyzx
- */
- #ifndef _ZX_FFT_H
- #define _ZX_FFT_H
- #include "../Headers/Global.h"
- #define TYPE_FFT_E float /* Type is the same with COMPLEX member */
- #ifndef PI
- #define PI (3.14159265f)
- #endif
- typedef COMPLEX TYPE_FFT; /* Define COMPLEX in Config.h */
- extern int fft(TYPE_FFT *x, int N);
- extern int ifft(TYPE_FFT *x, int N);
- #endif /* ZX_FFT_H_ */
/** zx_fft.h** Created on: 2013-8-5* Author: monkeyzx*/#ifndef _ZX_FFT_H
#define _ZX_FFT_H#include "../Headers/Global.h"#define TYPE_FFT_E float /* Type is the same with COMPLEX member */ #ifndef PI
#define PI (3.14159265f)
#endiftypedef COMPLEX TYPE_FFT; /* Define COMPLEX in Config.h */extern int fft(TYPE_FFT *x, int N);
extern int ifft(TYPE_FFT *x, int N);#endif /* ZX_FFT_H_ */
- /*
- * zx_fft.c
- *
- * Implementation of Fast Fourier Transform(FFT)
- * and reversal Fast Fourier Transform(IFFT)
- *
- * Created on: 2013-8-5
- * Author: monkeyzx
- *
- * TEST OK 2014.01.14
- * == 2014.01.14
- * Replace @BitReverse(x,x,N,M) by refrence to
- * <The Scientist and Engineer's Guide to Digital Signal Processing>
- */
- #include "zx_fft.h"
- /*
- * FFT Algorithm
- * === Inputs ===
- * x : complex numbers
- * N : nodes of FFT. @N should be power of 2, that is 2^(*)
- * === Output ===
- * the @x contains the result of FFT algorithm, so the original data
- * in @x is destroyed, please store them before using FFT.
- */
- int fft(TYPE_FFT *x, int N)
- {
- int i,j,l,k,ip;
- static int M = 0;
- static int le,le2;
- static TYPE_FFT_E sR,sI,tR,tI,uR,uI;
- M = (int)(log(N) / log(2));
- /*
- * bit reversal sorting
- */
- l = N / 2;
- j = l;
- //BitReverse(x,x,N,M);
- for (i=1; i<=N-2; i++) {
- if (i < j) {
- tR = x[j].real;
- tI = x[j].imag;
- x[j].real = x[i].real;
- x[j].imag = x[i].imag;
- x[i].real = tR;
- x[i].imag = tI;
- }
- k = l;
- while (k <= j) {
- j = j - k;
- k = k / 2;
- }
- j = j + k;
- }
- /*
- * For Loops
- */
- for (l=1; l<=M; l++) { /* loop for ceil{log2(N)} */
- le = (int)pow(2,l);
- le2 = (int)(le / 2);
- uR = 1;
- uI = 0;
- sR = cos(PI / le2);
- sI = -sin(PI / le2);
- for (j=1; j<=le2; j++) { /* loop for each sub DFT */
- //jm1 = j - 1;
- for (i=j-1; i<=N-1; i+=le) { /* loop for each butterfly */
- ip = i + le2;
- tR = x[ip].real * uR - x[ip].imag * uI;
- tI = x[ip].real * uI + x[ip].imag * uR;
- x[ip].real = x[i].real - tR;
- x[ip].imag = x[i].imag - tI;
- x[i].real += tR;
- x[i].imag += tI;
- } /* Next i */
- tR = uR;
- uR = tR * sR - uI * sI;
- uI = tR * sI + uI *sR;
- } /* Next j */
- } /* Next l */
- return 0;
- }
- /*
- * Inverse FFT Algorithm
- * === Inputs ===
- * x : complex numbers
- * N : nodes of FFT. @N should be power of 2, that is 2^(*)
- * === Output ===
- * the @x contains the result of FFT algorithm, so the original data
- * in @x is destroyed, please store them before using FFT.
- */
- int ifft(TYPE_FFT *x, int N)
- {
- int k = 0;
- for (k=0; k<=N-1; k++) {
- x[k].imag = -x[k].imag;
- }
- fft(x, N); /* using FFT */
- for (k=0; k<=N-1; k++) {
- x[k].real = x[k].real / N;
- x[k].imag = -x[k].imag / N;
- }
- return 0;
- }
/** zx_fft.c** Implementation of Fast Fourier Transform(FFT)* and reversal Fast Fourier Transform(IFFT)* * Created on: 2013-8-5* Author: monkeyzx** TEST OK 2014.01.14* == 2014.01.14* Replace @BitReverse(x,x,N,M) by refrence to * <The Scientist and Engineer's Guide to Digital Signal Processing>*/#include "zx_fft.h"/** FFT Algorithm* === Inputs ===* x : complex numbers* N : nodes of FFT. @N should be power of 2, that is 2^(*)* === Output ===* the @x contains the result of FFT algorithm, so the original data* in @x is destroyed, please store them before using FFT.*/
int fft(TYPE_FFT *x, int N)
{int i,j,l,k,ip;static int M = 0;static int le,le2;static TYPE_FFT_E sR,sI,tR,tI,uR,uI;M = (int)(log(N) / log(2));/** bit reversal sorting*/l = N / 2;j = l;//BitReverse(x,x,N,M);for (i=1; i<=N-2; i++) {if (i < j) {tR = x[j].real;tI = x[j].imag;x[j].real = x[i].real;x[j].imag = x[i].imag;x[i].real = tR;x[i].imag = tI;}k = l;while (k <= j) {j = j - k;k = k / 2;}j = j + k;}/** For Loops*/for (l=1; l<=M; l++) { /* loop for ceil{log2(N)} */le = (int)pow(2,l);le2 = (int)(le / 2);uR = 1;uI = 0;sR = cos(PI / le2);sI = -sin(PI / le2);for (j=1; j<=le2; j++) { /* loop for each sub DFT *///jm1 = j - 1;for (i=j-1; i<=N-1; i+=le) { /* loop for each butterfly */ip = i + le2;tR = x[ip].real * uR - x[ip].imag * uI;tI = x[ip].real * uI + x[ip].imag * uR;x[ip].real = x[i].real - tR;x[ip].imag = x[i].imag - tI;x[i].real += tR;x[i].imag += tI;} /* Next i */tR = uR;uR = tR * sR - uI * sI;uI = tR * sI + uI *sR;} /* Next j */} /* Next l */return 0;
}/** Inverse FFT Algorithm* === Inputs ===* x : complex numbers* N : nodes of FFT. @N should be power of 2, that is 2^(*)* === Output ===* the @x contains the result of FFT algorithm, so the original data* in @x is destroyed, please store them before using FFT.*/
int ifft(TYPE_FFT *x, int N)
{int k = 0;for (k=0; k<=N-1; k++) {x[k].imag = -x[k].imag;}fft(x, N); /* using FFT */for (k=0; k<=N-1; k++) {x[k].real = x[k].real / N;x[k].imag = -x[k].imag / N;}return 0;
}
另外,可能还需要您在其它头文件中定义COMPLEX的复数类型
- typedef struct {
- float real;
- float imag;
- } COMPLEX;
typedef struct {float real;float imag;
} COMPLEX;
=====================
增加:FFT频谱结果显示
=====================
- clc;
- clear;
- % Read a real audio signal
- [fname,pname]=uigetfile(...
- {'*.wav';'*.*'},...
- 'Input wav File');
- [x fs nbits] = wavread(fullfile(pname,fname));
- % Window
- % N = 1024;
- N = size(x,1);
- x = x(1:N, 1);
- % FFT
- y = fft(x);
- % 频率对称,取N/2
- y = y(1:N/2);
- % FFT频率与物理频率转换
- x1 = 1:N/2;
- x2 = (1:N/2)*fs/(N/2); % /1000表示KHz
- log_x2 = log2(x2);
- % plot
- figure,
- subplot(2,2,1);plot(x);
- xlabel('Time/N');ylabel('Mag');grid on
- title('原始信号');
- subplot(2,2,2);plot(x1, abs(y));
- xlabel('Freq/N');ylabel('Mag');grid on
- title('FFT信号/横坐标为点数');
- subplot(2,2,3);plot(x2,abs(y));
- xlabel('Freq/Hz');ylabel('Mag');grid on
- title('FFT信号/横坐标为物理频率');
- subplot(2,2,4);plot(log_x2,abs(y));
- xlabel('Freq/log2(Hz)');ylabel('Mag');grid on
- title('FFT信号/横坐标为物理频率取log');
- % 更常见是将幅值取log
- y = log2(abs(y));
- figure,
- plot(x2,y);
- xlabel('Freq/Hz');ylabel('Mag/log2');grid on
- title('幅值取log2');
clc;
clear;% Read a real audio signal
[fname,pname]=uigetfile(...{'*.wav';'*.*'},...'Input wav File');
[x fs nbits] = wavread(fullfile(pname,fname));% Window
% N = 1024;
N = size(x,1);
x = x(1:N, 1);% FFT
y = fft(x);
% 频率对称,取N/2
y = y(1:N/2);% FFT频率与物理频率转换
x1 = 1:N/2;
x2 = (1:N/2)*fs/(N/2); % /1000表示KHz
log_x2 = log2(x2);% plot
figure,
subplot(2,2,1);plot(x);
xlabel('Time/N');ylabel('Mag');grid on
title('原始信号');
subplot(2,2,2);plot(x1, abs(y));
xlabel('Freq/N');ylabel('Mag');grid on
title('FFT信号/横坐标为点数');
subplot(2,2,3);plot(x2,abs(y));
xlabel('Freq/Hz');ylabel('Mag');grid on
title('FFT信号/横坐标为物理频率');
subplot(2,2,4);plot(log_x2,abs(y));
xlabel('Freq/log2(Hz)');ylabel('Mag');grid on
title('FFT信号/横坐标为物理频率取log');% 更常见是将幅值取log
y = log2(abs(y));
figure,
plot(x2,y);
xlabel('Freq/Hz');ylabel('Mag/log2');grid on
title('幅值取log2');
最终部分优化的源代码我放在https://github.com/xiahouzuoxin/fft。
FFT算法的完整DSP实现相关推荐
- FFT算法的完整DSP实现(转)
源:FFT算法的完整DSP实现 傅里叶变换或者FFT的理论参考: [1] http://www.dspguide.com/ch12/2.htm The Scientist and Engineer's ...
- FFT算法的DSP实现
FFT算法移植到DSP的过程 一, 学习导向 为什么要学习5000系列DSP?我想这就像为什么要学习<信号与系统>,<模拟电路>一样,他不仅是理论性强的东西,更是直接具 ...
- JavaScript实现快速傅立叶变换FFT算法(附完整源码)
JavaScript实现快速傅立叶变换FFT算法(附完整源码) radianToDegree.js完整源代码 ComplexNumber.js完整源代码 bitLength.js完整源代码 fastF ...
- dsp实现快速傅里叶的C语言程序,DSP-快速傅立叶变换(FFT)算法实验
<DSP-快速傅立叶变换(FFT)算法实验>由会员分享,可在线阅读,更多相关<DSP-快速傅立叶变换(FFT)算法实验(10页珍藏版)>请在人人文库网上搜索. 1.中 南 大 ...
- 基于傅里叶变换的音频重采样算法 (附完整c代码)
前面有提到音频采样算法: WebRTC 音频采样算法 附完整C++示例代码 简洁明了的插值音频重采样算法例子 (附完整C代码) 近段时间有不少朋友给我写过邮件,说了一些他们使用的情况和问题. 坦白讲, ...
- c代码实现 ifft运算_fft算法c语言_matlab fft算法_ifft c语言
FFT快速算法C程序_工学_高等教育_教育专区.电子信息工程综合课程设计报告书 DSP 课程设计 报告 题学 目: 院: FFT 快速算法 C 程序 计算机与信息工程学院 09 ... fft算法代码 ...
- 快速傅氏变换之旅(五) 嵌入式中的FFT(最好选择DSP或FPGA)
欢迎大家提意见讨论. 转载请标明是引用于 http://blog.csdn.net/chenyujing1234 例子代码:(编译工具:VS2005) ==================理论==== ...
- 傅里叶变换 ~ 基 2 时间抽取 FFT 算法
文章目录 1.基2时间抽取FFT算法原理 2.基2时间抽取FFT算法流图 2.1.示例1 ~ 4点的序列表示成两个2点的DFT 2.2.示例2 ~ 8点的序列表示成两个2点的DFT 2.3.实例演示 ...
- 基于matlab和FFT算法实现信号频谱分析
系列文章目录 数字信号处理(DSP:Digital Signal Process)是电子通信领域非常重要的研究方向,博主汇总了数字信号处理(DSP)中常用的经典案例分析,主要基于算法分析.MATLAB ...
最新文章
- 《Windows网络与通信程序设计(第3版)》——1.4 网络应用程序设计基础
- FreeSWITCH IVR中lua调用并执行nodejs代码
- php依次替换src,如何在php中替换img中src内容
- JAVA使用正则表达式给字符串添加分隔符
- 【译】UNIVERSAL IMAGE LOADER. PART 3---ImageLoader详解
- Linux namespace之:network namespace
- java java 检查型异常_如何整合Java中的有效性检查和异常抛出?
- 想要酷炫大气的网页设计?这样做超吸睛
- BZOJ3240 NOI2013矩阵游戏(数论)
- MFC动态菜单全攻略
- window系统修复
- 全志android编译过程
- Android studio 录屏并制作gif
- python表格数据对比_python入门之对比两份excel表格数据
- 阿里开女性创业者大会 马云:世界因为女性而美好
- Python量化交易之预测茅台股票涨跌
- Linux_29_Linux-Vsftpd
- iOS开发:如何使用ShareSDK让APP快速拥有分享功能
- Jquery 模仿Baidu、Google收索时提示输入结果
- 拉勾教育 重学数据结构与算法 08 | 字符串:如何正确回答面试中高频考察的字符串匹配算法?