sfftools/sfftobmp3/tags/REL_3_1_4/doc
pschaefer 455d121464 2012-09-22 15:16:16 +00:00
..
getopt 2012-09-22 15:16:16 +00:00
changes 2012-09-22 15:16:16 +00:00
copying 2012-09-22 15:16:16 +00:00
credits 2012-09-22 15:16:16 +00:00
readme 2012-09-22 15:16:16 +00:00

readme

---------------
SFFTOBMP V3.1.4
---------------

Tool to convert "Structured Fax Files" into Windows bitmaps (BMP), 
Portable Bitmaps (PBM), JPEG Pictures (JPG) and single or multi-
paged TIFF Images.


SFF - Wotsit ?
--------------

SFF means 'Structured Fax File', intended by the CAPI interface to be 
used to save incoming faxes. Within Linux it's e.g. the AVM-B1 card 
which creates such files. 

See http://www.capi.org for specs of the CAPI interface.

This is where sfftobmp comes in.


Features
--------

Conversion of SFF files in Modified Huffman (MH) coding to Windows 
BMP (monochrome, uncompressed), PBM (Portable Bitmap with RAWBITS 
P4-Header), JPEG or TIFF files.

SFF files containing more than one page will be split into one 
output file per page, except for TIFF output which will be one multi-
paged file (however, even for TIFF there's an option to create 
singlepaged files).

Written in C++


Compilation
-----------

Only tested under Ubuntu Linux ('Gutsy') and Windows2000 SP4.

Used compilers:

gcc 4.1.3
Visual Studio Express

Under Linux, a simple call to './configure && make' will be sufficient 
most of the time if using GNU make and provided all needed libraries 
are installed (see below).

Under Windows, a Visual Studio workspace file is no longer supplied.
Please see the folder "win32" for a Makefile for Microsoft NMAKE to 
compile sfftobmp on the command line. Use

 nmake nodebug=1 /f makefile.vc
 
to build (run the VCVARSALL.BAT before to set all environment variables).

Alternatively, you can use the free MinGW C++ Compiler. 
A good (and also free) IDE is Dev-C++ from Bloodshed Software - you'll 
find a workspace file for this IDE in the distribution ("sfftobmp.dev").
See http://www.bloodshed.net/dev/devcpp.html. However, you will have to
update the paths to external libraries contained in this workspace.

To compile sfftobmp you need the following 3rd party libraries:

  - jpeglib
  - tifflib
  - libboost-filesystem (available since Boost 1.31.0).

On a Linux system these libraries can usually be installed via the package
manager. On a Debian system you need

 libjpeg62 and libjpeg62-dev
 libtiff4 and libtiff4-dev
 libboost-filesystem-dev

That's all.

On Windows, the sources are supplied in the "win32" folder, the makefiles
are already preconfigured for statically linkage with sfftobmp. 

For Dev-C++ pre-packaged versions of Tfflib and Jpeglib are available.
See http://www.bloodshed.net/dev/packages/imagelib.html. Boost libraries
have to be compiled by yourself, tough (they do provide MinGW support).

The 3rd party libraries can be downloaded at this locations:

 Boost libraries:  http://www.boost.org/
 Tifflib        :  http://www.remotesensing.org/libtiff/
 Jpeglib        :  http://www.ijg.org/


Usage
-----

  sfftobmp [options] INFILE1 [INFILE2 ..] [-o OUTSPEC]

  Options:
    -h or -help       Show help message and exit
    -v or -version    Show version and exit
    -b or -bmp        Output is one monochrome BMP file for each fax page
    -p or -pbm        Output is one Portable Bitmap file for each fax page
    -j or -jpg        Output is one JPEG file for each fax page
    -jNUM or -jpg=NUM Use jpeg quality of NUM percent (1..99)
    -T or -tifs       Output is one single-page TIFF file for each fax page
    -t or -tif        Output is one multi-paged TIFF file containing all pages
    -r or -keepres    Inhibit line doubling for low-res faxes
    -d or -keepdate   Keep date and time of input file for the output file(s)
    -q or -quiet      Don't display messages (except errors)
  
OUTSPEC is interpreted as:
   
 - a filename if only one input file is given.
 - a directory name if more than one input file is given.
  
If OUTSPEC is ommitted, the name of the output files will be 
derived from the input files and created in the same directory.
  
In case of TIFF output, you can specify the compression by using 
additional specifiers:
   
 -tr resp. -Tr : CCITT modified Huffman RLE
 -t4 resp. -T4 : CCITT Fax Class 4 compression
  
The default is CCITT Fax Class 3 compression.

For JPEG output you can optionally set the JPEG quality used by 
adding a two-digit number between 1 and 99 to the option. 

I.e.

-j35       or
-jpg=35

means JPEG output with 35% quality. If you don't give a quality, 
a default of 40% is used (mostly sufficient for b&w faxes).


OUTPUT ON STDOUT
----------------

Output on stdout is available for multipaged TIFF output 
only (option "-t"). Use "-" as output filename in this 
case:

 sfftobmp -t infile.sff -o - > out.tif

NOTE: Reading from stdin is not supported.

 
OUTPUT FILENAMES
----------------

The output filenames will be created by replacing the
extension of the input file (if it exists) by the extension 
of the target format. If the input file contains more than 
one page a three-digit number is added.

So, if the input is called "foo.sff" and has 3 pages,
the output files are named "foo_001.bmp", "foo_002.bmp"
and "foo_003.bmp" if you don't specify an OUTSPEC.

If you specify an OUTSPEC with an extension, say 
"sfftobmp -b foo.sff -o bar.out" the output filenames are 
called "bar_001.out", "bar_002.out" and "bar_003.out" - 
that is the given extension ".out" is preserved in this
case.

If you specify an OUTSPEC with no extension, the output
files will also carry no extension.


RESOLUTION HANDLING
-------------------

SffToBmp keeps resolution information intact. This means, if
you convert a fax with 203x98 dpi this resolution information is
written in the output file if it is supported by the output
format. This is JPEG, TIFF and BMP.

However, as many bitmap viewer applications are not considering
this information while displaying the bitmap, SffToBmp defaults
to double scanlines if a low-res fax (i.e. 203x98 dpi) is 
converted.

In case your image viewer does actually treat resolution information
correctly (your image will look stretched vertically in this case), 
you can suppress this scanline-doubling by specifying the "-r" option.


TODO
----

- speed up decoder with better table lookups
  (looking at GHz CPUs noadays, i doubt that this happes...) 
- support other non-intel architectures (64 Bits, different
  byte order). Sfftobmp will most likely fail on these.
  (well, no one complained till now)
- BMP format stores the scan lines bottom-up, that means
  the first line in the file is the last on the screen...
  Therefore the BMP target file will be filled from the end to
  the beginning. To do that, a seek beyond the end of file 
  is necessary at creation of the file. This may lead to problems
  on some platforms.


Last but not least...
---------------------

Should decoding of a SFF file fail:
Please send the SFF file uuencoded to me - if it isn't about
500 MB and/or contains top secret information ;).

Suggestions, extensions and patches welcome, flames will go 
to /dev/null.

I'm always pleased to hear that somebody is actually using my
software. If you can manage it, e-mail me a quick notice.

If you *really* like sfftobmp or depend on it, please look
into my wishlist:

http://www.amazon.de/gp/registry/wishlist/OVQ6LSYS4E4D/028-1843567-9543743

Thanks!

Contact:
--------
peter.schaefer@gmx.net
http://sfftools.sourceforge.net

$Id$