djvudigital - creates DjVu files from PS or PDF files.
djvudigital [options] inputfile
[outputfile]
This program creates a DjVu file from the PostScript
(.ps), GZipped PostScript (.ps.gz), Encapsulated
PostScript (.eps), or Portable Document Format
(.pdf) file inputfile.
The output file name is either given by argument outputfile
or generated by replacing the input file name suffixes by the DjVu suffix
(.djvu).
This program depends on a specific GhostScript driver. If your
GhostScript program does not provide this driver, please check
http://djvu.sourceforge.net/gsdjvu.html.
- --verbose,
-v
- Displays more informational messages while converting the file.
- --quiet,
-q
- Do not display informational messages while converting the file.
- --dpi=resolution
- Specify the desired resolution to resolution dots per inch. The
default is 300 dpi.
- --psrotate=angle
- Rotate the PostScript file by angle degrees clockwise. Only the
values 0, 90, 180, and 270 are supported. This
option only applies to PostScript files. PDF files are always converted
according to their native orientation.
- --epsf=disposition
- Specify how to handle Encapsulated PostScript files. Argument
disposition can take the values crop, fit, and
ignore. The default disposition crop creates a DjVu file
whose size matches the bounding box of the Encapsulated PostScript file.
Value fit rescales the graphics to the default page size. Value
ignore disables all Encapsulated PostScript specific code. This
option requires Ghostscript 7.07 or better.
- --exact-color
- Enables a more accurate rendering of the colors. This option requires
GhostScript 6.52 or better.
- --threshold=thres
- Specify a threshold for the foreground/background separation code.
Acceptable values of thres range from 0 to 100. Larger values place
more information into the foreground layer. The default threshold value is
80.
- --bg-subsample=sub
- Specify the background subsampling ratio. Argument sub must be an
integer between 1 and 6. The default value is 3.
- --bg-slices=n+...+n
- Specify the encoding quality of the background layer. The syntax for the
argument is similar to that described for the -slice option of
command c44. The default is 72+11+10+10.
- --fg-colors=ncolors
- Specify the maximum number of distinct colors in the foreground layer.
Argument ncolors can take integer values between 1 and 4000. The
default value is 256.
- --fg-image-colors=ncolors
- Specify the maximum number of distinct colors in an image for considering
encoding it into the foreground layer. Argument ncolors can take
integer values between 1 and 4000. The default value is 256.
- --words
- Extract the text from the PostScript code and incorporates this
information into the DjVu file. This option records the location of every
word.
- --lines
- Extract the text from the PostScript code and incorporates this
information into the DjVu file. This option saves a few bytes by only
recording the location of each line.
- --gsarg=arg1[,arg2,...,argN]
- Insert extra arguments on the GhostScript command line.
- --cseparg=arg1[,arg2,...,argN]
- Insert extra arguments on the command line of program csepdjvu or
msepdjvu.
- --poppler=keywords
- This option causes djvudigital to extract additional information
from PDF files using the tool pdftotext that comes bundled with the
Poppler library. Selected information is then added to the djvu file as a
postprocessing step. This option is ignored when the input file is not a
PDF file. Argument keywords is a comma separated list of keywords.
When this list contains keyword meta, the metadata extracted by
pdftotext is inserted into the djvu file. When this list contains
keyword text, the textual information extracted by pdftotext
is inserted into the djvu file, possibly replacing the information
gathered using the options --words or --lines. This is
useful for instance when a scanned PDF file contains a hidden text layer
that is not recognized by Ghostscript and therefore not passed to the
djvudigital backend.
- --sepfile
- Produces a separated data file instead of a DjVu file. Program
csepdjvu can then convert the separated data file into a DjVu
file.
- --check
- Display the names of the two auxiliary programs found by
djvudigital, namely a suitable ghostscript interpreter and a
suitable backend encoder. See the next two section for details.
- --dryrun
- Simply display the ghostscript command line generated by
djvudigital without running it. No output file is produced
- --help
- Display the manual page for djvudigital.
Program djvudigital internally relies on a specific
Ghostscript driver named djvusep. This driver analyzes the logical
structure of the sequence of PostScript rendering commands and decides to
execute each command into either the foreground or the background layer. The
GhostScript driver produces a separated data file that is then compressed
using the DjVuLibre program csepdjvu.
Before processing the input file, program djvudigital
searches a Ghostscript executable providing the djvusep driver. The
search starts with the file specified by the environment variable
GSDJVU and continues with command line executables named gs
and gsdjvu.
The DjVuLibre source code contains instruction to compile such a
GhostScript executable. More information can be obtained from
http://djvu.sourceforge.net/gsdjvu.html.
The output of the djvusep GhostScript driver must be
processed by the DjVuLibre program csepdjvu. This program can also be
replaced by the the proprietary Lizardtech program msepdjvu. Before
processing the input file, program djvudigital searches such an
executable. The search starts with the file specified by the environment
variable CSEPDJVU and continues with command line executables named
msepdjvu and csepdjvu.
The option --poppler=keywords relies on the tool
pdftotext that comes with the Poppler library and the tool
djvused that comes with djvulibre. Only recent versions of
pdftotext that accept the option -bbox are supported. Both
tools are searched by first trying the files specified by the environment
variables PDFTOTEXT and DJVUSED, and then trying executables
named pdftotext or djvused found along the shell executable
path.
The first version of this converter was written by Léon
Bottou <leonb@users.sourceforge.net> in AT&T Labs. The DjVuLibre
version is derived from code graciously released by Lizardtech in January
2004.
Program djvudigital can only process input files that
GhostScript can process properly.