diff --git a/README.rst b/README.rst
index 1d4a152..3038d5f 100644
--- a/README.rst
+++ b/README.rst
@@ -1,5 +1,5 @@
 =======================================
-syncstart(1) Version 1.1.1 \| syncstart
+syncstart(1) Version 1.1.2 \| syncstart
 =======================================
 
 SYNOPSIS
@@ -7,15 +7,16 @@ SYNOPSIS
 
 Command line help::
 
-    usage: syncstart [-h] [--version] [-v] [-b BEGIN] [-t TAKE] [-n] [-d] [-l LOWPASS] [-c] [-s] [-q] in1 in2
+    usage: syncstart [options] in1 in2
     
-    CLI interface to sync two media files using their audio or video streams.
-      ffmpeg needs to be available.
+    CLI program to compute timing offset (seconds) of media file 1 (in1)
+    in referenceto media file 2 (in2) using their audio or video streams.
+    ffmpeg is required.
       
     
     positional arguments:
-      in1                   First media file to sync with second.
-      in2                   Second media file to sync with first.
+      in1                   Offset media file.
+      in2                   Reference media file.
     
     options:
       -h, --help            show this help message and exit
@@ -44,9 +45,9 @@ The steps taken by ``syncstart``:
 - process and extract sample audio/video clips using ffmpeg with some default and optional filters
 - read the two clips into a 1D array and apply optional z-score normalization
 - compute offset via correlation using scipy ifft/fft
-- print ffmpeg/ffprobe output or optionally quiet that
-- show diagrams to allow MANUAL correction using ZOOM or optionally suppress that
-- print result
+- print ffmpeg/ffprobe output (optional)
+- show diagrams to allow MANUAL correction using ZOOM (optional)
+- print result as human readable, or print and return result as CSV
 
 MANUAL correction with ZOOM:
 
@@ -79,28 +80,44 @@ INSTALLATION
 
 To install for user only, do::
 
-   pip install --user syncstart
+    pip install --user syncstart
 
 Or activate a virtualenv and do::
 
-   pip install syncstart
+    pip install syncstart
+
+To make syncstart an executable in your PATH on Windows, do::
+
+    # Install pip if you don't have it
+    py -m ensurepip
+    # Upgrade and ensure pip prefers official wheels
+    py -m pip install --upgrade pip setuptools wheel
+    # Install/update pipx
+    py -m pip install --user --upgrade pipx
+    # Ensure python binaries are in PATH
+    py -m pipx ensurepath
+    pipx install syncstart
+    # Upgrade numpy inside pipx venv
+    pipx runpip syncstart install --upgrade numpy
+    # Test syncstart.exe by reading the version
+    syncstart --version
 
 EXAMPLES
 --------
 
 ::
 
-  # compute audio offset with default settings:
-  syncstart from_s10.m4a from_gopro.m4p
-
-  # compute audio offset using first 10 seconds with denoising, normalization and a 300 Hz lowpass filter:
-  syncstart video1.mp4 video2.mkv -t 10 -dnl 300
-
-  # compute video offset using first 20 seconds, don't show plots, only output final result:
-  syncstart video1.mp4 video2.mkv -vsq
-
-  # compute video offset using seconds 15 to 25 with denoising, cropping and normalization:
-  syncstart video1.mp4 video2.mkv -b 15 -t 10 -vdcn
+    # compute audio offset with default settings:
+    syncstart from_s10.m4a from_gopro.m4p
+    
+    # compute audio offset using first 10 seconds with denoising, normalization and a 300 Hz lowpass filter:
+    syncstart -t 10 -dnl 300 video1.mp4 video2.mkv
+    
+    # compute video offset using first 20 seconds, don\[aq]t show plots, only output final result:
+    syncstart -vsq video1.mp4 video2.mkv
+    
+    # compute video offset using seconds 15 to 25 with denoising, cropping and normalization:
+    syncstart -b 15 -t 10 -vdcn video1.mp4 video2.mkv
 
 License
 -------
diff --git a/syncstart.py b/syncstart.py
index 4011aee..b0130ed 100755
--- a/syncstart.py
+++ b/syncstart.py
@@ -263,8 +263,7 @@ def on_zoom(event_ax):
     else:
         return ff,noff
   else:
-    return ff, toff
-
+    return ff, round(toff, 3)
 
 def cli_parser(**ka):
   import argparse
@@ -277,11 +276,11 @@ def cli_parser(**ka):
   if 'in1' not in ka:
     parser.add_argument(
       'in1',
-      help='First media file to sync with second.')
+      help='Offset media file.')
   if 'in2' not in ka:
     parser.add_argument(
       'in2',
-      help='Second media file to sync with first.')
+      help='Reference media file.')
   if 'video' not in ka:
     parser.add_argument(
       '-v','--video',
@@ -380,7 +379,28 @@ def file_offset(**ka):
   """Sync two media files using their audio or video streams.
   ffmpeg is required.
   """
+  ls1 = len(s1)
+  ls2 = len(s2)
+  padsize = ls1+ls2+1
+  padsize = 2**(int(np.log(padsize)/np.log(2))+1)
+  s1pad = np.zeros(padsize)
+  s1pad[:ls1] = s1
+  s2pad = np.zeros(padsize)
+  s2pad[:ls2] = s2
+  corr = scipy.fft.ifft(scipy.fft.fft(s1pad)*np.conj(scipy.fft.fft(s2pad)))
+  ca = np.absolute(corr)
+  xmax = np.argmax(ca)
+  offset_samples = xmax
+  if xmax > padsize // 2:
+    offset_samples = xmax-padsize
+  return offset_samples, ca
+
 
+def file_offset(**ka):
+  """CLI program to compute timing offset (seconds) of media file 1 (in1)
+in referenceto media file 2 (in2) using their audio or video streams.
+  ffmpeg is required.
+  """
   parser = cli_parser(**ka)
   args = parser.parse_args().__dict__
   ka.update(args)
@@ -398,26 +418,21 @@ def file_offset(**ka):
   padsize,xmax,correlation = padsize_xmax_correlation(s1,s2)
   if show: show1(sr,correlation,title='Correlation',v=xmax/sr)
   sync_text = """
-==============================================================================
-%s needs 'ffmpeg -ss %s' cut to get in sync
-==============================================================================
+============
+The first input is offset %s' seconds.
+============
 """
-  if xmax > padsize // 2:
-    if show:
-      file,offset = show2(sr,s1,s2,-(padsize-xmax),in1,in2)
-    else:
-      file,offset = in2,(padsize-xmax)/sr
+  if show:
+    file, offset_seconds = show2(sr, s1, s2, offset_samples, in1, in2)
   else:
-    if show:
-      file,offset = show2(sr,s1,s2,xmax,in1,in2)
-    else:
-      file,offset = in1,xmax/sr
+    file,offset_seconds = in1,round(offset_samples/sr, 3)
   if not quiet: #default
-    print(sync_text%(file,offset))
+    print(sync_text%(offset_seconds))
   else: #quiet
     ## print csv: file_to_advance,seconds_to_advance
-    print("%s,%s"%(file,offset))
-  return file,offset
+    print("%s,%s"%(file,offset_seconds))
+  return file,offset_seconds
+
 
 main = file_offset
 if __name__ == '__main__':