Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/jasonacox/sfw2jpg

Convert SFW (Seattle FilmWorks) image files to JPG Format (sfwjpg pwpjpg flipjpeg)
https://github.com/jasonacox/sfw2jpg

jpeg-images seattle-filmworks sfw

Last synced: about 1 month ago
JSON representation

Convert SFW (Seattle FilmWorks) image files to JPG Format (sfwjpg pwpjpg flipjpeg)

Awesome Lists containing this project

README 

        
#

# details.txt

# 

# Copyright (c) 1997-1999  Everett Lipman

#

# This document is free software; you can redistribute it and/or

# modify it under the terms of the GNU General Public License

# as published by the Free Software Foundation; either version 2

# of the License, or (at your option) any later version.

# 

# This document is distributed in the hope that it will be useful,

# but WITHOUT ANY WARRANTY; without even the implied warranty of

# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

# GNU General Public License for more details.

# 

# The GNU General Public License is available at

#

#    http://www.fsf.org/copyleft/gpl.html

#

# Alternatively, you can write to the 

#

#    Free Software Foundation, Inc.

#    59 Temple Place - Suite 330

#    Boston, MA  02111-1307, USA

#



#

# 9 May 1999

#



This file contains instructions for the compilation of, and

information about the function of, the programs sfwjpg, pwpjpg, and

flipjpeg.



NOTE: The Independent JPEG group's jpegtran program now performs

      the function of flipjpeg, which was written before the present

      version of jpegtran became available.



      Seattle FilmWorks recently began scanning pictures at

      resolutions other than 640x480.  This change affects all

      resolutions referred to throughout this document.



Compilation:



  sfwjpg and pwpjpg



    sfwjpg.c and pwpjpg.c are standalone ANSI C programs which should

    compile without any modifications under UNIX.  Compilation under

    Windows with Microsoft Visual C++ requires that the strcasecmp()

    function, which occurs in both programs, be replaced with

    strcmp().



  flipjpeg - See NOTE above



    flipjpeg.c is an ANSI C program which uses the Independent JPEG

    Group's (IJG) free JPEG software, which you can find at

    ftp://ftp.uu.net/graphics/jpeg.  flipjpeg.c uses standard input

    and output, and has only been tested under UNIX.  It is probably

    possible to compile flipjpeg.c on another platform with a few

    modifications, such as changing the input and output files.  See

    the IJG software documentation for more information.



    In order to compile flipjpeg.c under UNIX, unpack the IJG

    software and place flipjpeg.c in the same directory.  After

    running 'configure' to create the Makefile for the IJG software,

    add the following lines to Makefile:



    FJOBJECTS= flipjpeg.o rdswitch.o cdjpeg.o



    flipjpeg: $(FJOBJECTS) libjpeg.a

        $(LN) $(LDFLAGS) -o flipjpeg $(FJOBJECTS) libjpeg.a $(LDLIBS)



    then type



      make flipjpeg



Program Operation



  This section contains a technical description of the function of

  sfwjpg, pwpjpg, and flipjpeg.  If you need more information about

  JPEG compression or the JFIF file format, you can consult



  THE JPEG STILL IMAGE DATA COMPRESSION STANDARD

  by William Pennebaker and Joan Mitchell

  Van Nostrand Reinhold, 1993



  The information presented here about .sfw files has been deduced

  through examination of a number of example files, and may not be

  complete.  None of this information has been supplied or confirmed

  by Seattle FilmWorks.



  sfwjpg



    sfwjpg converts a file in the .sfw format to the standard JPEG

    File Interchange Format (JFIF).  This is not terribly difficult,

    since a .sfw file is really a JFIF file with a few

    modifications.  Here are the steps in the conversion, as

    performed by sfwjpg:



    1.  Find the start of the JFIF data.

        In a .sfw file, this is delineated by the four byte

        sequence



        0xff 0xc8 0xff 0xd0



        this sequence must be changed to 



        0xff 0xd8 0xff 0xe0



        These four bytes become the start of the JFIF (.jpg) file,

        all prior data having been discarded.



    2.  Place JFIF identifier in the proper place.

        The following 7 byte string must appear after the first six

        bytes, which are the 4 initial bytes from step 1, plus two

        others, which indicate the length of the data preceding

        the next marker:



        "JFIF\0\001\0"  Note: no implicit trailing null in this string.



    3.  Fix the other JPEG markers in the file.

        Two byte sequences beginning with 0xff, such as those in

        step 1, are called markers, and indicate what type

        of data follows in the JPEG stream or JFIF file.  In the

        .sfw file, the following markers must be converted back

        to those specified in the JPEG standard.  The .sfw

        markers are on the left, and the corresponding JPEG

        markers are on the right:



        0xff 0xa0  ->  0xff 0xc0

        0xff 0xa4  ->  0xff 0xc4

        0xff 0xc8  ->  0xff 0xd8   (done in step 1)

        0xff 0xc9  ->  0xff 0xd9

        0xff 0xca  ->  0xff 0xda

        0xff 0xcb  ->  0xff 0xdb

        0xff 0xd0  ->  0xff 0xe0   (done in step 1)



    4.  Remove trailing data after the 0xff 0xd9 (0xc9 before correction)

        marker near the end of the file.

    

    5.  Insert the Huffman coding table if it is not present in the

        .sfw file.  This table, along with its marker, can be found in

        sfwjpg.c.  The table should be inserted just before the

        0xff 0xda (0xca before correction) marker.



    Once these steps have been carried out, you will have a JFIF file

    containing your image.  The image will still need to be flipped

    vertically.



  pwpjpg

  

    A .pwp file consists of a group of .sfw files (typically all of

    the pictures from a roll of film) which have had some additional

    information prepended to them (such as the file name and length).

    pwpjpg reads the length and file name information, and then passes

    each .sfw file to the same routine used in sfwjpg, which converts

    the file to JFIF and writes it to disk.



    The beginning of the first filename is located at byte 0x14 in the

    .pwp file, with byte 0x00 being the first byte in the file.  The

    start of the .sfw file corresponding to that filename can be

    found by searching for "SFW94A".  The length of the .sfw file is

    stored in little-endian order.  If the char pointer pwpptr

    points to the start of the "SFW94A" string, the three bytes

    of the file length can be found at pwpptr-0x0A, pwpptr-0x0B,

    and pwpptr-0x0C.  The byte at pwpptr-0x0A is the most

    significant.  The next filename begins right after the end

    of the previous .sfw file.  See the pwpjpg.c code for

    further details.



  flipjpeg



    flipjpeg does a lossless inversion of the image in a JFIF file.



    The JPEG compression process consists of a number of steps.

    My (probably incomplete and possibly incorrect) understanding 

    of this process is as follows:



    1.  The image to be compressed is converted to a

        representation using the desired components, for

        example, Y, Cb and Cr instead of R, G and B.



    2.  Each component is broken up into 8x8 pixel blocks, and

        a two dimensional discrete cosine transform (DCT) is

        performed on each block of pixels.  Sometimes at this

        stage, various components are reduced in resolution.  In

        .sfw files, the Y component is kept at 640x480, and the

        Cb and Cr components are reduced to 320x240.



    4.  The DCT coefficients are "quantized" (reduced in numerical,

        not spatial, resolution) in order to reduce the amount of

        information in the image in such a way that the appearance

        of the image will not be noticeably affected.

 

    5.  The zero frequency coefficients for each block are written

        so that they represent the change in the coefficient from the

        previous block instead of the actual value of the coefficient.



    6.  The coefficients are compressed using Huffman coding, and the

        resulting data are written out along with some markers to

        delineate coding tables, etc.



    In order to do a lossless inversion, flipjpeg operates on raw DCT

    coefficients.  First, it uses the Independent JPEG Group's

    (excellent) software to read the DCT coefficients for each

    component into an array.  It then goes through each array,

    flipping the sign of every DCT coefficient which corresponds to a

    basis function that is asymmetric in the vertical direction.

    This operation has the effect of vertically flipping each of the

    8 pixel high horizontal stripes in the image.  The coefficient

    block rows are finally rearranged in the proper order, and the

    file is written out, again using the IJG software.



    This procedure works fine on .sfw-derived JPEG files, since

    Seattle FilmWorks scans images at 640x480 resolution, and 480

    happens to be a multiple of 16.  When the image height is not a

    multiple of 8, the remaining pixel rows in the last block are

    filled with the same values as the last row in the image.  If the

    components of the image are not sampled at the same resolution

    (and perhaps sometimes even if they are) the different components

    will be offset from each other in the flipped image.  This can be

    seen if flipjpeg is run on testimg.jpg from the IJG software

    distribution.