diff --git a/Makefile b/Makefile
index de261c2..91acf21 100644
--- a/Makefile
+++ b/Makefile
@@ -1,5 +1,7 @@
+SPEEDTESTFILE=/dev/shm/kissfft_speedtest
+SPEEDTESTNSAMPS=1000
 
-all: kiss_fft_s kiss_fft_f kiss_fft_d freqpeak tones testsig
+all: kiss_fft_s kiss_fft_f kiss_fft_d
 
 kiss_fft_s: kiss_fft.h kiss_fft.c
 	gcc -Wall -O3 -o kiss_fft_s -DFIXED_POINT -DFFT_UTIL kiss_fft.c -lm
@@ -10,27 +12,17 @@ kiss_fft_f: kiss_fft.h kiss_fft.c
 kiss_fft_d: kiss_fft.h kiss_fft.c
 	gcc -Wall -O3 -o kiss_fft_d -Dkiss_fft_scalar=double -DFFT_UTIL kiss_fft.c -lm
 	
-freqpeak: kiss_fft.h kiss_fft.c freqpeak.c
-	gcc -Wall -O3 -o freqpeak freqpeak.c kiss_fft.c -lm
-
-testsig: testsig.c
-	gcc -Wall -O3 -o testsig testsig.c -lm
-
-tones: tones.c
-	gcc -Wall -O3 -o tones tones.c -lm
-	
 clean:
-	rm -f kiss_fft_s kiss_fft_f kiss_fft_d *~ fftin.dat fftout.dat \
-	freqpeak testsig tones
+	rm -f kiss_fft_s kiss_fft_f kiss_fft_d *~ fftin.dat fftout.dat $(SPEEDTESTFILE)
 
 test: all
 	./test.oct
 
-speedtest: /dev/shm/junk kiss_fft_f
-	time ./kiss_fft_f < /dev/shm/junk > /dev/null
+speedf: kiss_fft_f  $(SPEEDTESTFILE)
+	time ./kiss_fft_f < $(SPEEDTESTFILE) > /dev/null
 
-/dev/shm/junk:
-	dd if=/dev/urandom bs=8192 count=1000 of=/dev/shm/junk
+$(SPEEDTESTFILE):
+	dd if=/dev/zero bs=8192 count=$(SPEEDTESTNSAMPS) of=$(SPEEDTESTFILE)
 
 tarball: clean
 	tar -czf kiss_fft.tar.gz .
diff --git a/README b/README
index a7644a4..ef72078 100644
--- a/README
+++ b/README
@@ -12,10 +12,14 @@ USAGE:
 
         void * cfg = kiss_fft_alloc( nfft ,inverse_fft );
         while ...
-            ... 
+            ... // put kth sample in cx_buf_in_out[k].r and cx_buf_in_out[k].i
             kiss_fft( cfg , cx_buf_in_out );
-            ...
+            ... // transformed 
         free(cfg);
+        
+    Note: frequency-domain data is stored from dc to 2pi.
+    so cx_buf_in_out[0] is the dc bin of the FFT
+    and cx_buf_in_out[nfft/2] is the Nyquist bin
 
     Declarations are in "kiss_fft.h", along with a brief description of the 
 two functions you'll need to use. Code definitions are in kiss_fft.c, along 
@@ -53,8 +57,8 @@ last bit of performance.
 PERFORMANCE:
     (on Athlon XP 2100+, with gcc 2.96, optimization O3, float data type)
 
-    Kiss performed 1000 1024-pt ffts in 136 ms. This translates to 7.5 Msamples/s.
-    Just for comparison, it took md5sum 160 ms to process the same amount of data
+    Kiss performed 1000 1024-pt ffts in 110 ms of cpu time (132ms real time). 
+    For comparison, it took md5sum 160ms cputime to process the same amount of data
 
 DO NOT:
     ... use Kiss if you need the absolute fastest fft in the world
@@ -67,13 +71,16 @@ UNDER THE HOOD:
 addressing is corrected as the last step in the transform.  No scaling is done.
 
 LICENSE:
-    BSD, see COPYING for details.
+    BSD, see COPYING for details. Basically, "free to use, give credit where due, no guarantees"
 
 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.
+    *) Add sample code for parallel ffts (stereo) packed into re,im components of time sequence.
+    *) Add simple windowing function, e.g. Hamming : w(i)=.54-.46*cos(2pi*i/(n-1))
+    *) Could mixed-radix FFTs be made simple enough to stand by the KISS principle?
+    *) 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.
 
 AUTHOR:
     Mark Borgerding
-    mark@borgerding.net
+    Mark@Borgerding.net
diff --git a/freqpeak.c b/freqpeak.c
deleted file mode 100644
index cd9e2ff..0000000
--- a/freqpeak.c
+++ /dev/null
@@ -1,87 +0,0 @@
-#include <stdio.h>
-#include <string.h>
-#include <memory.h>
-#include <malloc.h>
-#include <stdio.h>
-#include <math.h>
-#include "kiss_fft.h"
-
-#define NFFT 1024
-int main(int argc, char ** argv)
-{
-    int k;
-    void * st;
-    float fs=44100;
-
-    short sampsin[2*NFFT];
-    float lmag2[NFFT/2];
-    float rmag2[NFFT/2];
-    int peakr=0,peakl=0;
-    int removedc=1;
-
-    kiss_fft_cpx cbuf[NFFT];
-    int nbufs=0;
-
-    st = kiss_fft_alloc(NFFT,0);
-
-    memset( lmag2 , 0 , sizeof(lmag2) );
-    memset( rmag2 , 0 , sizeof(rmag2) );
-
-    while ( fread( sampsin , sizeof(short) * 2*NFFT, 1 , stdin ) == 1 ) {
-
-        //perform two ffts in parallel by packing the left&right channels into the real and imaginary
-        for (k=0;k<NFFT;++k) {
-            cbuf[k].r = sampsin[2*k];
-            cbuf[k].i = sampsin[2*k+1];
-        }
-
-        if (removedc) {
-            float dcr=0,dci=0;
-            for (k=0;k<NFFT;++k){
-                dcr += cbuf[k].r;
-                dci += cbuf[k].i;
-            }
-            dcr /= NFFT;
-            dci /= NFFT;
-
-            for (k=0;k<NFFT;++k) {
-                cbuf[k].r -= dcr;
-                cbuf[k].i -= dci;
-            }
-        }
-
-        // perform the fft on the L+R packed buffer
-        kiss_fft( st , cbuf );
-
-        // get the half-symmetric FFTs for the Left and Right Channels
-        for (k=0;k<NFFT/2;++k) {
-            int k2 = (NFFT-k)%NFFT;
-            kiss_fft_cpx r,l;
-
-            r.r = (cbuf[k].r + cbuf[k2].r) * 0.5;
-            r.i = (cbuf[k].i - cbuf[k2].i) * 0.5;
-
-            l.r = (cbuf[k].i + cbuf[k2].i) * 0.5;
-            l.i = (cbuf[k].r - cbuf[k2].r) * 0.5;
-
-            rmag2[k] += r.r * r.r + r.i * r.i;
-            lmag2[k] += l.r * l.r + l.i * l.i;
-        }
-        ++nbufs;
-    }
-    free(st);
-
-    for (k=0;k<NFFT/2;++k) {
-        if (rmag2[peakr] < rmag2[k])
-            peakr = k;
-        if (lmag2[peakl] < lmag2[k])
-            peakl = k;
-    }
-
-    printf("%d buffers\n",nbufs);
-
-    printf("peak frequency R:%.3fdB @ %.1f Hz\n",10*log10(rmag2[peakr]) , peakr*fs/NFFT);
-    printf("               L:%.3fdB @ %.1f Hz\n",10*log10(lmag2[peakl]) , peakl*fs/NFFT);
-
-    return 0;
-}
diff --git a/kiss_fft.c b/kiss_fft.c
index 0685619..ecb1987 100644
--- a/kiss_fft.c
+++ b/kiss_fft.c
@@ -140,7 +140,6 @@ void undo_bit_rev( kiss_fft_cpx * f, const int * swap_indices,int nswap)
 }
 
 
-#define OPT1
 // the heart of the fft
 static
 void fft_work(int N,kiss_fft_cpx *f,const kiss_fft_cpx * twid,int twid_step)
@@ -149,23 +148,10 @@ void fft_work(int N,kiss_fft_cpx *f,const kiss_fft_cpx * twid,int twid_step)
     {   // declare automatic variables in here to reduce stack usage
         int n;
         kiss_fft_cpx csum,cdiff;
-#ifdef OPT1
-        const kiss_fft_cpx * cur_twid = twid+twid_step;
-        kiss_fft_cpx *f1 = f;
-        kiss_fft_cpx *f2 = f+N;
-
-        C_SUB( cdiff, *f1 , *f2 );
-        C_ADD( csum,  *f1 , *f2 );
-        *f1++ = csum;
-        *f2++ = cdiff;
-
-        for (n=1;n<N;++n)
-#else
         const kiss_fft_cpx * cur_twid = twid;
         kiss_fft_cpx *f1 = f;
         kiss_fft_cpx *f2 = f+N;
         for (n=0;n<N;++n)
-#endif
         {
             C_ADD( csum,  *f1 , *f2 );
             C_SUB( cdiff, *f1 , *f2 );
diff --git a/testsig.c b/testsig.c
deleted file mode 100644
index 7b2c007..0000000
--- a/testsig.c
+++ /dev/null
@@ -1,32 +0,0 @@
-#include <stdio.h>
-#include <math.h>
-#include <stdlib.h>
-
-int usage()
-{
-    fprintf(stderr,"usage:testsig nsamps\n");
-    exit(1);
-    return 1;
-}
-
-double randphase()
-{
-    return (double)rand()*2*3.14159/RAND_MAX;
-}
-
-int main(int argc, char ** argv)
-{
-    float samps[2];
-    int nsamps;
-
-    if (argc != 2)
-        return usage();
-    nsamps = atoi( argv[1] );
-    
-    while (nsamps-- > 0) {
-        samps[0]=sin( randphase() );
-        samps[1]=sin( randphase() );
-        fwrite(samps,sizeof(samps),1,stdout);
-    }
-    return 0;
-}
diff --git a/tones.c b/tones.c
deleted file mode 100644
index 43f9f89..0000000
--- a/tones.c
+++ /dev/null
@@ -1,35 +0,0 @@
-
-#include <stdio.h>
-#include <string.h>
-#include <memory.h>
-#include <malloc.h>
-#include <stdio.h>
-#include <math.h>
-#include "kiss_fft.h"
-
-#define PI 3.14159
-
-int main(int argc, char ** argv)
-{
-    int k;
-    float fs=44100;
-
-    float fr=1,fl=300;
-
-    float th[2] = {0,0};
-    float thinc[2] = {2*PI*fr/fs,2*PI*fl/fs };
-
-    while (1){
-        for (k=0;k<2;++k){
-            short s;
-            th[k] += thinc[k];
-            if (th[k] > 2*PI){
-                th[k] -= 2*PI;
-            }
-            s=(short)32767*cos( th[k] );
-            fwrite(&s,sizeof(s),1,stdout);
-        }
-    }
-
-    return 0;
-}