summaryrefslogtreecommitdiffstats
path: root/libFDK/src/fft_rad2.cpp
diff options
context:
space:
mode:
Diffstat (limited to 'libFDK/src/fft_rad2.cpp')
-rw-r--r--libFDK/src/fft_rad2.cpp450
1 files changed, 450 insertions, 0 deletions
diff --git a/libFDK/src/fft_rad2.cpp b/libFDK/src/fft_rad2.cpp
new file mode 100644
index 0000000..80467ec
--- /dev/null
+++ b/libFDK/src/fft_rad2.cpp
@@ -0,0 +1,450 @@
+/*************************** Fraunhofer IIS FDK Tools **********************
+
+ (C) Copyright Fraunhofer IIS (2003)
+ All Rights Reserved
+
+ Please be advised that this software and/or program delivery is
+ Confidential Information of Fraunhofer and subject to and covered by the
+
+ Fraunhofer IIS Software Evaluation Agreement
+ between Google Inc. and Fraunhofer
+ effective and in full force since March 1, 2012.
+
+ You may use this software and/or program only under the terms and
+ conditions described in the above mentioned Fraunhofer IIS Software
+ Evaluation Agreement. Any other and/or further use requires a separate agreement.
+
+
+ $Id$
+ Author(s): M. Lohwasser, M. Gayer
+ Description:
+
+ This software and/or program is protected by copyright law and international
+ treaties. Any reproduction or distribution of this software and/or program,
+ or any portion of it, may result in severe civil and criminal penalties, and
+ will be prosecuted to the maximum extent possible under law.
+
+******************************************************************************/
+
+#include "fft_rad2.h"
+
+#include "scramble.h"
+
+#define __FFT_RAD2_CPP__
+
+#if defined(__arm__) /* cppp replaced: elif */
+#include "arm/fft_rad2_arm.cpp"
+
+#elif defined(__GNUC__) && defined(__mips__) && defined(__mips_dsp) /* cppp replaced: elif */
+#include "mips/fft_rad2_mips.cpp"
+
+#endif
+
+
+
+/*****************************************************************************
+
+ functionname: dit_fft (analysis)
+ description: dit-tukey-algorithm
+ scrambles data at entry
+ i.e. loop is made with scrambled data
+ returns:
+ input:
+ output:
+
+*****************************************************************************/
+
+#ifndef FUNCTION_dit_fft
+
+void dit_fft(FIXP_DBL *x, const INT ldn, const FIXP_STP *trigdata, const INT trigDataSize)
+{
+ const INT n=1<<ldn;
+ INT trigstep,i,ldm;
+
+ scramble(x,n);
+ /*
+ * 1+2 stage radix 4
+ */
+
+ for (i=0;i<n*2;i+=8)
+ {
+ FIXP_DBL a00, a10, a20, a30;
+ a00 = (x[i + 0] + x[i + 2])>>1; /* Re A + Re B */
+ a10 = (x[i + 4] + x[i + 6])>>1; /* Re C + Re D */
+ a20 = (x[i + 1] + x[i + 3])>>1; /* Im A + Im B */
+ a30 = (x[i + 5] + x[i + 7])>>1; /* Im C + Im D */
+
+ x[i + 0] = a00 + a10; /* Re A' = Re A + Re B + Re C + Re D */
+ x[i + 4] = a00 - a10; /* Re C' = Re A + Re B - Re C - Re D */
+ x[i + 1] = a20 + a30; /* Im A' = Im A + Im B + Im C + Im D */
+ x[i + 5] = a20 - a30; /* Im C' = Im A + Im B - Im C - Im D */
+
+ a00 = a00 - x[i + 2]; /* Re A - Re B */
+ a10 = a10 - x[i + 6]; /* Re C - Re D */
+ a20 = a20 - x[i + 3]; /* Im A - Im B */
+ a30 = a30 - x[i + 7]; /* Im C - Im D */
+
+ x[i + 2] = a00 + a30; /* Re B' = Re A - Re B + Im C - Im D */
+ x[i + 6] = a00 - a30; /* Re D' = Re A - Re B - Im C + Im D */
+ x[i + 3] = a20 - a10; /* Im B' = Im A - Im B - Re C + Re D */
+ x[i + 7] = a20 + a10; /* Im D' = Im A - Im B + Re C - Re D */
+ }
+
+ for(ldm=3; ldm<=ldn; ++ldm)
+ {
+ INT m=(1<<ldm);
+ INT mh=(m>>1);
+ INT j,r;
+
+ trigstep=((trigDataSize << 2)>>ldm);
+
+ FDK_ASSERT(trigstep > 0);
+
+ /* Do first iteration with c=1.0 and s=0.0 separately to avoid loosing to much precision.
+ Beware: The impact on the overal FFT precision is rather large. */
+ {
+ j = 0;
+
+ for(r=0; r<n; r+=m)
+ {
+ INT t1 = (r+j)<<1;
+ INT t2 = t1 + (mh<<1);
+ FIXP_DBL vr,vi,ur,ui;
+
+ //cplxMultDiv2(&vi, &vr, x[t2+1], x[t2], (FIXP_SGL)1.0, (FIXP_SGL)0.0);
+ vi = x[t2+1]>>1;
+ vr = x[t2]>>1;
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui-vi;
+
+ t1 += mh;
+ t2 = t1+(mh<<1);
+
+ //cplxMultDiv2(&vr, &vi, x[t2+1], x[t2], (FIXP_SGL)1.0, (FIXP_SGL)0.0);
+ vr = x[t2+1]>>1;
+ vi = x[t2]>>1;
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui-vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui+vi;
+ }
+ }
+ for(j=1; j<mh/4; ++j)
+ {
+ FIXP_STP cs;
+
+ cs = trigdata[j*trigstep];
+
+ for(r=0; r<n; r+=m)
+ {
+ INT t1 = (r+j)<<1;
+ INT t2 = t1 + (mh<<1);
+ FIXP_DBL vr,vi,ur,ui;
+
+ cplxMultDiv2(&vi, &vr, x[t2+1], x[t2], cs);
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui-vi;
+
+ t1 += mh;
+ t2 = t1+(mh<<1);
+
+ cplxMultDiv2(&vr, &vi, x[t2+1], x[t2], cs);
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui-vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui+vi;
+
+ /* Same as above but for t1,t2 with j>mh/4 and thus cs swapped */
+ t1 = (r+mh/2-j)<<1;
+ t2 = t1 + (mh<<1);
+
+ cplxMultDiv2(&vi, &vr, x[t2], x[t2+1], cs);
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui-vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui+vi;
+
+ t1 += mh;
+ t2 = t1+(mh<<1);
+
+ cplxMultDiv2(&vr, &vi, x[t2], x[t2+1], cs);
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur-vr;
+ x[t1+1] = ui-vi;
+
+ x[t2] = ur+vr;
+ x[t2+1] = ui+vi;
+ }
+ }
+ {
+ j = mh/4;
+
+ for(r=0; r<n; r+=m)
+ {
+ INT t1 = (r+j)<<1;
+ INT t2 = t1 + (mh<<1);
+ FIXP_DBL vr,vi,ur,ui;
+
+ cplxMultDiv2(&vi, &vr, x[t2+1], x[t2], STC(0x5a82799a), STC(0x5a82799a));
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui-vi;
+
+ t1 += mh;
+ t2 = t1+(mh<<1);
+
+ cplxMultDiv2(&vr, &vi, x[t2+1], x[t2], STC(0x5a82799a), STC(0x5a82799a));
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui-vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui+vi;
+ }
+ }
+ }
+}
+#endif
+
+
+/*****************************************************************************
+
+ functionname: dit_ifft (synthesis)
+ description: dit-tukey-algorithm
+ scrambles data at entry
+ i.e. loop is made with scrambled data
+ returns:
+ input:
+ output:
+
+*****************************************************************************/
+
+#if !defined(FUNCTION_dit_ifft)
+void dit_ifft(FIXP_DBL *x, const INT ldn, const FIXP_STP *trigdata, const INT trigDataSize)
+{
+ const INT n=1<<ldn;
+ INT trigstep,i,ldm;
+
+ scramble(x,n);
+
+ /*
+ 1+2 stage radix 4
+ */
+
+ for (i=0;i<n*2;i+=8)
+ {
+ FIXP_DBL a0, a1, a2, a3, a00, a10, a20, a30;
+
+ a00 = (x[i + 0] + x[i + 2])>>1; /* Re A + Re B */
+ a10 = (x[i + 4] + x[i + 6])>>1; /* Re C + Re D */
+ a20 = (x[i + 1] + x[i + 3])>>1; /* Im A + Im B */
+ a30 = (x[i + 5] + x[i + 7])>>1; /* Im C + Im D */
+ a0 = (x[i + 0] - x[i + 2])>>1; /* Re A - Re B */
+ a2 = (x[i + 4] - x[i + 6])>>1; /* Re C - Re D */
+ a3 = (x[i + 1] - x[i + 3])>>1; /* Im A - Im B */
+ a1 = (x[i + 5] - x[i + 7])>>1; /* Im C - Im D */
+
+ x[i + 0] = a00 + a10; /* Re A' = Re A + Re B + Re C + Re D */
+ x[i + 4] = a00 - a10; /* Re C' = Re A + Re B - Re C - Re D */
+ x[i + 1] = a20 + a30; /* Im A' = Im A + Im B + Im C + Im D */
+ x[i + 5] = a20 - a30; /* Im C' = Im A + Im B - Im C - Im D */
+ x[i + 2] = a0 - a1; /* Re B' = Re A - Re B - Im C + Im D */
+ x[i + 6] = a0 + a1; /* Re D' = Re A - Re B + Im C - Im D */
+ x[i + 3] = a3 + a2; /* Im B' = Im A - Im B + Re C - Re D */
+ x[i + 7] = a3 - a2; /* Im D' = Im A - Im B - Re C + Re D */
+ }
+
+ for(ldm=3; ldm<=ldn; ++ldm)
+ {
+ const INT m=(1<<ldm);
+ const INT mh=(m>>1);
+
+ INT j,r;
+
+ trigstep=((trigDataSize << 2)>>ldm);
+
+ {
+ j = 0;
+
+ for(r=0; r<n; r+=m)
+ {
+ INT t1 = (r+j)<<1;
+ INT t2 = t1 + (mh<<1);
+ FIXP_DBL vr,vi,ur,ui;
+
+ //cplxMultDiv2(&vr, &vi, x[t2], x[t2+1], FL2FXCONST_SGL(1.0), (FIXP_SGL)0.0);
+ vi = x[t2+1]>>1;
+ vr = x[t2]>>1;
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui-vi;
+
+ t1 += mh;
+ t2 = t1+(mh<<1);
+
+ //cplxMultDiv2(&vi, &vr, x[t2], x[t2+1], FL2FXCONST_SGL(1.0), FL2FXCONST_SGL(0.0));
+ vr = x[t2+1]>>1;
+ vi = x[t2]>>1;
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur-vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur+vr;
+ x[t2+1] = ui-vi;
+ }
+ }
+ for(j=1; j<mh/4; ++j)
+ {
+ FIXP_STP cs;
+
+ cs = trigdata[j*trigstep];
+
+ for(r=0; r<n; r+=m)
+ {
+ INT t1 = (r+j)<<1;
+ INT t2 = t1 + (mh<<1);
+ FIXP_DBL vr,vi,ur,ui;
+
+ cplxMultDiv2(&vr, &vi, x[t2], x[t2+1], cs);
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui-vi;
+
+ t1 += mh;
+ t2 = t1+(mh<<1);
+
+ cplxMultDiv2(&vi, &vr, x[t2], x[t2+1], cs);
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur-vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur+vr;
+ x[t2+1] = ui-vi;
+
+ /* Same as above but for t1,t2 with j>mh/4 and thus cs swapped */
+ t1 = (r+mh/2-j)<<1;
+ t2 = t1 + (mh<<1);
+
+ cplxMultDiv2(&vr, &vi, x[t2+1], x[t2], cs);
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur-vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur+vr;
+ x[t2+1] = ui-vi;
+
+ t1 += mh;
+ t2 = t1+(mh<<1);
+
+ cplxMultDiv2(&vi, &vr, x[t2+1], x[t2], cs);
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur-vr;
+ x[t1+1] = ui-vi;
+
+ x[t2] = ur+vr;
+ x[t2+1] = ui+vi;
+ }
+ }
+ {
+ j = mh/4;
+ for(r=0; r<n; r+=m)
+ {
+ INT t1 = (r+mh/2-j)<<1;
+ INT t2 = t1 + (mh<<1);
+ FIXP_DBL vr,vi,ur,ui;
+
+ cplxMultDiv2(&vr, &vi, x[t2], x[t2+1], STC(0x5a82799a), STC(0x5a82799a));
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur+vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur-vr;
+ x[t2+1] = ui-vi;
+
+ t1 += mh;
+ t2 = t1+(mh<<1);
+
+ cplxMultDiv2(&vi, &vr, x[t2], x[t2+1], STC(0x5a82799a), STC(0x5a82799a));
+
+ ur = x[t1]>>1;
+ ui = x[t1+1]>>1;
+
+ x[t1] = ur-vr;
+ x[t1+1] = ui+vi;
+
+ x[t2] = ur+vr;
+ x[t2+1] = ui-vi;
+ }
+ }
+ }
+}
+#endif
+