luke
/
ngspice
mirror of https://git.code.sf.net/p/ngspice/ngspice
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Make fft scaling independent from rounding behaviour for odd data. 

Add a time slot to fft span.
This commit is contained in:
geraldfournier 2026-03-20 13:37:06 +01:00 committed by Holger Vogt
parent b3a78b02cc
commit 65e5ccb7bb
2 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
src/frontend/com_fft.c
View File

@ -59,10 +59,11 @@ com_fft(wordlist *wl)
length = (plot_cur->pl_scale)->v_length;
time = (plot_cur->pl_scale)->v_realdata;
span = time[length-1] - time[0];
span = time[length-1] - time[0] + time[length-1] - time[length-2];
#ifdef HAVE_LIBFFTW3
fpts = length/2 + 1;
scale = ((double)length)/2.0;
#else
/* size of fft input vector is power of two and larger or equal than spice vector */
N = 1;
@ -72,6 +73,7 @@ com_fft(wordlist *wl)
M++;
}
fpts = N/2 + 1;
scale = ((double)N)/2;
#endif
win = TMALLOC(double, length);
@ -178,7 +180,6 @@ com_fft(wordlist *wl)
fftw_execute(plan_forward);
scale = (double) fpts - 1.0;
fdvec[i][0].cx_real = out[0][0]/scale/2.0;
fdvec[i][0].cx_imag = 0.0;
for (j = 1; j < fpts; j++) {
@ -215,7 +216,6 @@ com_fft(wordlist *wl)
rffts(in, M, 1);
fftFree();
scale = (double) fpts - 1.0;
/* Re(x[0]), Re(x[N/2]), Re(x[1]), Im(x[1]), Re(x[2]), Im(x[2]), ... Re(x[N/2-1]), Im(x[N/2-1]). */
fdvec[i][0].cx_real = in[0]/scale/2.0;
fdvec[i][0].cx_imag = 0.0;

4
src/maths/cmaths/cmath4.c
View File

@ -774,7 +774,7 @@ cx_fft(void *data, short int type, int length, int *newlength, short int *newtyp
fftw_execute(plan_forward);
scale = (double) fpts - 1.0;
scale = ((double)length)/2.0;
outdata[0].cx_real = out[0][0]/scale/2.0;
outdata[0].cx_imag = 0.0;
for (i = 1; i < fpts; i++) {
@ -803,7 +803,7 @@ cx_fft(void *data, short int type, int length, int *newlength, short int *newtyp
rffts(datax, M, 1);
fftFree();
scale = (double) fpts - 1.0;
scale = ((double)N)/2;
/* Re(x[0]), Re(x[N/2]), Re(x[1]), Im(x[1]), Re(x[2]), Im(x[2]), ... Re(x[N/2-1]), Im(x[N/2-1]). */
outdata[0].cx_real = datax[0]/scale/2.0;
outdata[0].cx_imag = 0.0;