⭐ 欢迎来到虫虫下载站! | 📦 资源下载 📁 资源专辑 ℹ️ 关于我们
⭐ 虫虫下载站

📄 fft.c

📁 mp3文件格式与wav文件格式的音频文件转换工具
💻 C
字号:
#include <string.h>#include "types.h"#include "l3psy.h"#include "layer3.h"/***************************************************************************** ************************** Start of Subroutines ***************************** *****************************************************************************//***************************************************************************** * FFT computes fast fourier transform of BLKSIZE samples of data            * *   uses decimation-in-frequency algorithm described in "Digital            * *   Signal Processing" by Oppenheim and Schafer, refer to pages 304         * *   (flow graph) and 330-332 (Fortran program in problem 5)                 * *   to get the inverse fft, change line 20 from                             * *                 w_imag[L] = -sin(PI/le1);                                 * *                          to                                               * *                 w_imag[L] = sin(PI/le1);                                  * *                                                                           * *   required constants:                                                     * *         #define      PI          3.14159265358979                         * *         #define      BLKSIZE     1024                                     * *         #define      LOGBLKSIZE  10                                       * *         #define      BLKSIZE_S   256                                      * *         #define      LOGBLKSIZE_S 8                                       * *                                                                           * *****************************************************************************/#define      BLKSIZE_S   256#define      LOGBLKSIZE_S 8void fft(x_real,x_imag, energy, phi, N)float x_real[BLKSIZE], x_imag[BLKSIZE], energy[BLKSIZE], phi[BLKSIZE];int	N;{ int     M,MM1; static int     init=0; int     NV2, NM1, MP; static double  w_real[2][LOGBLKSIZE], w_imag[2][LOGBLKSIZE]; int            i,j,k,L; int            ip, le,le1; double         t_real, t_imag, u_real, u_imag; if(init==0) {    memset((char *) w_real, 0, sizeof(w_real));  /* preset statics to 0 */    memset((char *) w_imag, 0, sizeof(w_imag));  /* preset statics to 0 */    M = LOGBLKSIZE;    for(L=0; L<M; L++){       le = 1 << (M-L);       le1 = le >> 1;       w_real[0][L] = cos(PI/le1);       w_imag[0][L] = -sin(PI/le1);    }              M = LOGBLKSIZE_S;    for(L=0; L<M; L++){       le = 1 << (M-L);       le1 = le >> 1;       w_real[1][L] = cos(PI/le1);       w_imag[1][L] = -sin(PI/le1);    }              init++; } switch(N) {	case BLKSIZE:			M = LOGBLKSIZE;			MP = 0;			break;	case BLKSIZE_S:			M = LOGBLKSIZE_S;			MP = 1;			break;	default:	printf("Error: Bad FFT Size in subs.c\n");			exit(-1); } MM1 = M-1; NV2 = N >> 1; NM1 = N - 1; for(L=0; L<MM1; L++){    le = 1 << (M-L);    le1 = le >> 1;    u_real = 1;    u_imag = 0;    for(j=0; j<le1; j++){       for(i=j; i<N; i+=le){          ip = i + le1;          t_real = x_real[i] + x_real[ip];          t_imag = x_imag[i] + x_imag[ip];          x_real[ip] = x_real[i] - x_real[ip];          x_imag[ip] = x_imag[i] - x_imag[ip];          x_real[i] = t_real;          x_imag[i] = t_imag;          t_real = x_real[ip];          x_real[ip] = x_real[ip]*u_real - x_imag[ip]*u_imag;          x_imag[ip] = x_imag[ip]*u_real + t_real*u_imag;       }       t_real = u_real;       u_real = u_real*w_real[MP][L] - u_imag*w_imag[MP][L];       u_imag = u_imag*w_real[MP][L] + t_real*w_imag[MP][L];    } } /* special case: L = M-1; all Wn = 1 */ for(i=0; i<N; i+=2){    ip = i + 1;    t_real = x_real[i] + x_real[ip];    t_imag = x_imag[i] + x_imag[ip];    x_real[ip] = x_real[i] - x_real[ip];    x_imag[ip] = x_imag[i] - x_imag[ip];    x_real[i] = t_real;    x_imag[i] = t_imag;    energy[i] = x_real[i]*x_real[i] + x_imag[i]*x_imag[i];    if(energy[i] <= 0.0005){phi[i] = 0;energy[i] = 0.0005;}    else phi[i] = atan2((double) x_imag[i],(double) x_real[i]);    energy[ip] = x_real[ip]*x_real[ip] + x_imag[ip]*x_imag[ip];    if(energy[ip] == 0)phi[ip] = 0;    else phi[ip] = atan2((double) x_imag[ip],(double) x_real[ip]); } /* this section reorders the data to the correct ordering */ j = 0; for(i=0; i<NM1; i++){    if(i<j){/* use this section only if you need the FFT in complex number form * * (and in the correct ordering)                                    */       t_real = x_real[j];       t_imag = x_imag[j];       x_real[j] = x_real[i];       x_imag[j] = x_imag[i];       x_real[i] = t_real;       x_imag[i] = t_imag;/* reorder the energy and phase, phi                                        */       t_real = energy[j];       energy[j] = energy[i];       energy[i] = t_real;       t_real = phi[j];       phi[j] = phi[i];       phi[i] = t_real;    }    k=NV2;    while(k<=j){       j = j-k;       k = k >> 1;    }    j = j+k; }}

⌨️ 快捷键说明

复制代码 Ctrl + C
搜索代码 Ctrl + F
全屏模式 F11
切换主题 Ctrl + Shift + D
显示快捷键 ?
增大字号 Ctrl + =
减小字号 Ctrl + -