Legacy ImageMagick Discussions Archive

Hello,

I want to convert a postscript file (.PS) into a PDF file. I'm using the following command line:

convert -limit memory 1GiB -limit map 2GiB -define registry:temporary-path="c:\tmp" -density 300x300 -page A4 -colors 2 +dither -monochrome -compress group4  input.ps output.pdf

The problem is, that the size of the pdf is 5,04cm x 7,13cm (reported by Acrobat reader and PDF XChange Viewer). Expected is 21cm x 29,7cm.

For testing, I changed the output from PDF to TIF:

Code: Select all

convert -limit memory 1GiB -limit map 2GiB -define registry:temporary-path="c:\tmp" -density 300x300 -page A4 -colors 2 +dither -monochrome -compress group4  input.ps output.tif

The resulting picture is 2479 pixels x 3508 pixels. This is correct for A4 size at 300 dpi.

What is the cause? I'm using ImageMagick 6.8.9-9 Q16 x86 2014-10-19.

Best regards

akapuma

Code: Select all

convert -limit memory 1GiB -limit map 2GiB -define registry:temporary-path="c:\tmp" -density 300x300 -page A4 -colors 2 +dither -monochrome -compress group4  input.ps output.pdf

The syntax is wrong. It should be: input-file, processing, output file, like this:

Code: Select all

convert -limit memory 1GiB -limit map 2GiB -define registry:temporary-path="c:\tmp" input.ps -density 300x300 -page A4 -colors 2 +dither -monochrome -compress group4 output.pdf

Remove "-density 300x300". I think the size will then be A4.

The syntax is wrong. It should be: input-file, processing, output file, like this:

In this case, the result is wrong. For testing, you can use this test picture:
http://anders.geekhouse.no/Testpage/

Remove "-density 300x300". I think the size will then be A4.

Yes, this works, output is A4.
Problem: output is only 595x842 pixels = 72dpi.
I want to have 2479x3508 pixels = 300dpi

Best regards

akapuma

Hmm...

Why do you even use ImageMagick for this job in the first place?!?

IM cannot handle PS input directly. It employs Ghostscript as its delegate to convert the PS into a raster format it understands. You can observe this yourself by adding

Code: Select all

-verbose

to beginning of the commandline. It will show you a line like this:

Code: Select all

[ghostscript library] -q -dQUIET -dSAFER -dBATCH -dNOPAUSE -dNOPROMPT -dMaxBitmap=500000000 -dAlignToPixels=0 \
    -dGridFitTT=2 "-sDEVICE=pngalpha" -dTextAlphaBits=4 -dGraphicsAlphaBits=4 "-r72x72" -g192x192  \
    "-sOutputFile=/tmp/magick-18532WfuSEjmDC1iH%d" "-f/tmp/magick-18532ERyUj24PgJBq" "-f/tmp/magick-185327SZZPqvlFLP9"

So only after Ghostscript converted the PDF page to a PNG with an alpha channel ImageMagick will ever get hold of the result. When IM produces the output PDF, its pages will only contain raster data where before there were fonts and vector graphics.

Is this what you want?

If ImageMagick's convert does work at all, it means you have Ghostscript installed on your system anyway. So you could also go from PS -> PDF directly:

Code: Select all

gswin64c.exe -o output.pdf -sDEVICE=pdfwrite testpage.ps

Is the reason for using ImageMagick: you want to remove the colors from the PS and get a grayscale PDF? This can be achieved with using Ghostscript directly too:

Code: Select all

gswin64c.exe ^
  -o grayscale.pdf ^
  -sDEVICE=pdfwrite ^
  -sColorConversionStrategy=Gray ^
  -sProcessColorModel=DeviceGray ^
  -dOverrideICC ^
  -dCompatibilityLevel=1.4 ^
   c:/path/to/input.ps-or-pdf

Why do you even use ImageMagick for this job in the first place?!?

I want to have a PDF with the following requirements:

- no fonts, no vector graphics, bitmap graphics only
- Because the size of a bitmap is very high, a good compression is needed. Very effective is fax compression group 4.
- The fax compression needs a monochrome picture (1 bit per pixel).
- The monochrome compression should be done without dithering (looks better).

Your hint to gswin is very interesting! I will reed the option in the next days. Thank you.

Best regards

akapuma

To convert an A4 (color) PostScript into a grayscale PDF where each page contains only bitmaps with TIFF G4 (a.k.a. "fax") compression, and to have 300 ppi resolution per page, use these 2 commands:

Code: Select all

   gs -o testpage.tif -sDEVICE=tiffg4 -r300 testpage.ps

   tiff2pdf -p A4 -F testpage.tif > testpage.tif.pdf

If your final PDF should contain metadata for creator software, author, title, subject and keywords, insert the following parameters into the second command as needed:

Code: Select all

   -c "Creator software name"
   -a "Author name"
   -t "Document title"
   -s "Document subject"
   -k "Keyword1, Keyword2, Keyword3, ..."

To check the embedded images (their width, height, resolutions) use this command:
pdfimages -list testpage.tif.pdf

You should see:

Code: Select all

   page   num  type   width height color comp bpc  enc interp  object ID x-ppi y-ppi size ratio
   --------------------------------------------------------------------------------------------
      1     0 image    2550  3300  gray    1   1  ccitt  no         7  0   309   309  358K  35%

To control exactly the number of pixels in width and height of the TIFF, change the first command (Ghostscript):

Code: Select all

   gs -o testpage.tif -sDEVICE=tiffg4 -g2479x3508 testpage.ps

akapuma wrote:For testing, you can use this test picture:
http://anders.geekhouse.no/Testpage/

That image is square. If you want it to fill an A4 page, it will need stretching.

With IM, you can use "-density" before the input file to get a certain number of pixels. Then use another "-density" after the input file to change the density without resampling the pixels.

No, it's not square. It IS A4. Not sure what you base your statement about its squareness on.

According to Ghostscript, it's real BoundingBox (that is, the area, where ink would be deposited by a printer) is this:

Code: Select all

gs -q -o - -sDEVICE=bbox testpage.ps
%%BoundingBox: -1 -1 596 843
%%HiResBoundingBox: -0.008930 -0.008930 595.008966 842.008966

According to some (unofficial) PostScript standard if a size specification deviates by +5pt or -5pt from the print device's supported page size, it should still be deemed valid. (This originates from the PostScript Language and Reference, 3rd Ed, where the exact behaviour of a device is controlled by the /Policies <</PageSize N>> dictionary for cases where a requested media size cannot be matched exactly, but is within the -5pt -- +5pt range...)

When I do this:

Code: Select all

convert testpage.ps testpage.png

I get a center cropped region of the original that is square 192x192. I suspect that is what snibgo is suggesting by square.

IM see this for verbose information

Image: testpage.ps
Format: PS (PostScript)
Mime type: application/postscript
Class: DirectClass
Geometry: 192x192+0+0
Resolution: 72x72
Print size: 2.66667x2.66667
Units: Undefined
Type: TrueColorAlpha
Endianess: Undefined
Colorspace: sRGB
Depth: 16/8-bit
Page geometry: 192x192+0+0

You may have to use ghostscript standalone to avoid this cropping. I do not know where in the ps file data it says to crop.

pipitas wrote:No, it's not square. It IS A4. Not sure what you base your statement about its squareness on.

Ah, thanks. I was looking at the line "%%BoundingBox: 210 285 402 477". But I don't know much about PS.

The %%BoundingBox statement you found in the testpage.ps was the one coming from Einstein's picture, which is embedded in the main PS as an Encapsulated PostScript file. It is not at all valid and irrelevant for the overall main PS.

fmw42 wrote:IM see this for verbose information

Image: testpage.ps
Format: PS (PostScript)
Mime type: application/postscript
Class: DirectClass
Geometry: 192x192+0+0
Resolution: 72x72
Print size: 2.66667x2.66667
Units: Undefined
Type: TrueColorAlpha
Endianess: Undefined
Colorspace: sRGB
Depth: 16/8-bit
Page geometry: 192x192+0+0

Funnily, my version of IM reports this PS as a PNG file!

Code: Select all

identify -verbose testpage.ps | head -n 12

Image: testpage.ps
  Format: PNG (Portable Network Graphics)
  Mime type: image/png
  Class: DirectClass
  Geometry: 192x192+0+0
  Resolution: 72x72
  Print size: 2.66667x2.66667
  Units: Undefined
  Type: TrueColorAlpha
  Endianess: Undefined
  Colorspace: sRGB
  Depth: 16/8-bit

Something is seriously wrong with IM here.

It seems that it parses the entire file for occurences of '%%BoundingBox', takes the values found there (in this case: '210 285 402 477') and calculates from it the 'geometry' it reports.

This is plainly wrong. This procedure is only ever valid for a standalone EPS file, but never for a general PS file which may contain EPS files used as page elements (like here).

Funnily, my version of IM reports this PS as a PNG file!

I am using IM 6.8.9.10 Q16 Mac OSX.

Perhaps it has to do with the file on your image hosting server or the way it gets downloaded. If you download your file from the hosting service, do you still get PNG?

fmw42 wrote:
Funnily, my version of IM reports this PS as a PNG file!

6.8.9-8 Q16 Mac OSX here, installed via MacPorts.

pipitas wrote:
fmw42 wrote:
Funnily, my version of IM reports this PS as a PNG file!
6.8.9-8 Q16 Mac OSX here, installed via MacPorts.

I install my delegates from MacPorts, but install IM from source manually. see viewtopic.php?f=1&t=21502&p=88202&hilit ... rts#p88202

Legacy ImageMagick Discussions Archive

PDF is too small (in cm)

PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)

Re: PDF is too small (in cm)