From 71bf5965b87ced1da31705e0f3fec26b17dbc9f1 Mon Sep 17 00:00:00 2001
From: Mark Borgerding <mark@borgerding.net>
Date: Tue, 24 Feb 2004 02:21:03 +0000
Subject: [PATCH] very minor tweaks to radix 3 code

---
 CHANGELOG  | 10 ++++++++++
 Makefile   |  2 +-
 README     |  6 +++++-
 kiss_fft.c | 48 +++++++++++++++++++++++-------------------------
 4 files changed, 39 insertions(+), 27 deletions(-)

diff --git a/CHANGELOG b/CHANGELOG
index 6d74116..83d147b 100644
--- a/CHANGELOG
+++ b/CHANGELOG
@@ -1,3 +1,13 @@
+1.2 (Feb 23, 2004)
+    interface change -- cfg object is forward declaration of struct instead of void*
+    This maintains type saftey and lets the compiler warn/error about stupid mistakes.
+            (prompted by suggestion from Erik de Castro Lopo)
+
+    small speed improvements
+
+    added psdpng.c -- sample utility that will create png spectrum "waterfalls" from an input file
+        ( not terribly useful yet)
+
 1.1.1 (Feb 1, 2004 )
     minor bug fix -- only affects odd rank, in-place, multi-dimensional FFTs
 
diff --git a/Makefile b/Makefile
index 3fe8240..8df4c9a 100644
--- a/Makefile
+++ b/Makefile
@@ -1,4 +1,4 @@
-KFVER=111
+KFVER=120
 
 DISTDIR=kiss_fft_v$(KFVER)
 TARBALL=kiss_fft_v$(KFVER).tar.gz
diff --git a/README b/README
index ae61193..b7de350 100644
--- a/README
+++ b/README
@@ -12,7 +12,7 @@ USAGE:
 
         #include "kiss_fft.h"
 
-        void * cfg = kiss_fft_alloc( nfft ,inverse_fft );
+        kiss_fft_cfg cfg = kiss_fft_alloc( nfft ,inverse_fft );
 
         while ...
         
@@ -38,6 +38,7 @@ You can do other cool stuff with the extras you'll find in tools/
     * arbitrary dimension FFTs (complex only currently, apologies to Steve DeKorte -- mebbe next time )
     * real FFTs
     * fast convolution filtering
+    * spectrum image creation
 
 The core fft and most tools/ code can be compiled to use float, double 
 or 16bit short samples. The default is float.
@@ -101,6 +102,9 @@ TODO:
     *) Make the fixed point scaling and bit shifts more easily configurable.
     *) Document/revisit the input/output fft scaling
     *) See if the fixed point code can be optimized a little without adding complexity.
+    *) Make doc describing the overlap (tail) scrap fast convolution filtering in kiss_fastfir.c
+    *) Test all the ./tools/ code with fixed point (kiss_fastfir.c doesn't work, maybe others)
+
 
 AUTHOR:
     Mark Borgerding
diff --git a/kiss_fft.c b/kiss_fft.c
index 7627e23..2c506d5 100644
--- a/kiss_fft.c
+++ b/kiss_fft.c
@@ -26,7 +26,7 @@ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
 
 static void kf_bfly2(
         kiss_fft_cpx * Fout,
-        const int fstride,
+        const size_t fstride,
         const kiss_fft_cfg st,
         int m
         )
@@ -49,7 +49,7 @@ static void kf_bfly2(
 
 static void kf_bfly4(
         kiss_fft_cpx * Fout,
-        const int fstride,
+        const size_t fstride,
         const kiss_fft_cfg st,
         const size_t m
         )
@@ -96,53 +96,51 @@ static void kf_bfly4(
 
 static void kf_bfly3(
          kiss_fft_cpx * Fout,
-         const int fstride,
+         const size_t fstride,
          const kiss_fft_cfg st,
-         int m
+         size_t m
          )
 {
-     kiss_fft_cpx *Fout0,*Fout1,*Fout2;
+     size_t k=m;
+     const size_t m2 = 2*m;
      kiss_fft_cpx *tw1,*tw2;
      kiss_fft_cpx scratch[5];
      kiss_fft_cpx epi3;
      epi3 = st->twiddles[fstride*m];
 
-     Fout0=Fout;
-     Fout1=Fout0+m;
-     Fout2=Fout0+2*m;
      tw1=tw2=st->twiddles;
 
      do{
-         C_FIXDIV(*Fout0,3); C_FIXDIV(*Fout1,3); C_FIXDIV(*Fout2,3);
+         C_FIXDIV(*Fout,3); C_FIXDIV(Fout[m],3); C_FIXDIV(Fout[m2],3);
 
-         C_MUL(scratch[1],*Fout1 , *tw1);
-         C_MUL(scratch[2],*Fout2 , *tw2);
+         C_MUL(scratch[1],Fout[m] , *tw1);
+         C_MUL(scratch[2],Fout[m2] , *tw2);
 
          C_ADD(scratch[3],scratch[1],scratch[2]);
          C_SUB(scratch[0],scratch[1],scratch[2]);
+         tw1 += fstride;
+         tw2 += fstride*2;
 
-         Fout1->r = Fout0->r - scratch[3].r/2;
-         Fout1->i = Fout0->i - scratch[3].i/2;
+         Fout[m].r = Fout->r - scratch[3].r/2;
+         Fout[m].i = Fout->i - scratch[3].i/2;
 
          C_MULBYSCALAR( scratch[0] , epi3.i );
 
-         C_ADDTO(*Fout0,scratch[3]);
+         C_ADDTO(*Fout,scratch[3]);
 
-         Fout2->r = Fout1->r + scratch[0].i;
-         Fout2->i = Fout1->i - scratch[0].r;
+         Fout[m2].r = Fout[m].r + scratch[0].i;
+         Fout[m2].i = Fout[m].i - scratch[0].r;
 
-         Fout1->r -= scratch[0].i;
-         Fout1->i += scratch[0].r;
+         Fout[m].r -= scratch[0].i;
+         Fout[m].i += scratch[0].r;
 
-         ++Fout0;++Fout1;++Fout2;
-         tw1 += fstride;
-         tw2 += fstride*2;
-     }while(--m);
+         ++Fout;
+     }while(--k);
 }
 
 static void kf_bfly5(
         kiss_fft_cpx * Fout,
-        const int fstride,
+        const size_t fstride,
         const kiss_fft_cfg st,
         int m
         )
@@ -204,7 +202,7 @@ static void kf_bfly5(
 /* perform the butterfly for one stage of a mixed radix FFT */
 static void kf_bfly_generic(
         kiss_fft_cpx * Fout,
-        const int fstride,
+        const size_t fstride,
         const kiss_fft_cfg st,
         int m,
         int p
@@ -243,7 +241,7 @@ static
 void kf_work(
         kiss_fft_cpx * Fout,
         const kiss_fft_cpx * f,
-        const int fstride,
+        const size_t fstride,
         int in_stride,
         int * factors,
         const kiss_fft_cfg st