Convert a PDF to greyscale on the command line in FLOSS?

Question

I have a colour PDF file, and I'm going to print it out and then photocopy it in black and white. I'd like to know what it's like in B&W before photocopying it. Is it possible to 'greyscale' a PDF on the command line using free software? I'm using Ubuntu 9.10.

Solution 1

ImageMagick can do this.

convert -colorspace GRAY color.pdf gray.pdf

via this email

Solution 2

Better:

gs \
 -sOutputFile=output.pdf \
 -sDEVICE=pdfwrite \
 -sColorConversionStrategy=Gray \
 -dProcessColorModel=/DeviceGray \
 -dCompatibilityLevel=1.4 \
 -dNOPAUSE \
 -dBATCH \
 input.pdf

Solution 3

Here’s a little script which in addition to the grayscale conversion can concatenate multiple input files. To use the script, put the following lines in a file, e.g. "convert2gray.sh"

#!/bin/bash
gs -sOutputFile=converted.pdf -sDEVICE=pdfwrite -sColorConversionStrategy=Gray -dProcessColorModel=/DeviceGray -dCompatibiltyLevel=1.4 -dNOPAUSE -dBATCH $@

and make it executable

chmod +x convert2gray.sh

Then

./convert2gray.sh input1.pdf input2.pdf … lastinput.pdf

will produce a single PDF "converted.pdf", which contains all pages from the input files converted to grayscale.

I had to print out mutliple files all in grayscale and found this the easiest way, since you can print out everything after inpection with one command.

Comments

[tdc]

Agreed, this gives much better results than convert, but sometimes it rotates the pdf which is a bit annoying!

[tdc]

Just realised you can disable that with -dAutoRotatePages=/None

[Archie]

I just ran this command on a 58MB PDF that was already greyscale (came from a scanner) and the resulting output was 10MB and looked exactly the same. Nice!

[Johannes Weiss]

That significantly reduces quality. @goyinux' solution is better.

[ixe013]

Works well on Windows, too! Just remove the `\` and put everything on the same line.

[m000]

Convert will actually rasterize the contents of the pdf. So unless the pdf already encapsulates only raster images (e.g. a scanned document), this approach is a big no-no.

[burtek]

Unless you use -density 400 -quality 100 parameters - that works well

[jjmerelo]

In fact, this fails with this error GPL Ghostscript 9.10: Unable to convert color space to Gray, reverting strategy to LeaveColorUnchanged.

[saeedgnu]

Works great. But output image size is almost 3x for my file. I have a complicated vector-drawing colorful file with 3.5 MB size that became 10 MB!

[Joshua Grosso Reinstate CMs]

Really, is there anything ImageMagick can't do? :)

[user1338062]

FWIW, pdftk can concatenate PDFs, too.

[Stanimir Stoyanov]

-density 400 -quality 100 creates HUGE files. +1 for @goyinux's solution.

[rew]

The big files print great (only using black toner) while the original and gs-converted file use all toners and create a vague mess on my printer.

[Mpunkt Moeniac]

What's the reason for convert to reduce the quality of rasterized images? I tried using the options convert -density 300 -quality 100, in order to prevent this behaviour of convert. But the quality of the output image is still much worse than the quality of the input image... so are there options for convert, that prevent the encapsulated images from quality loss, when converting pdf'-files encapsulating rasterized images?

[xaratustra]

This is really fantastic. Perfect for scanned documents for subsequent print.

[loved.by.Jesus]

Another problem with convert is that OCRed pdf files lose the text layer. That is, one loses the OCR. While the method with gs keeps OCR.

[loved.by.Jesus]

Some people complains that the size of the grayscale image is not lower than the original. If, besides turning the image grayscale, you want to compress it, add the following option -dPDFSETTINGS=/ebook . Instead of /ebook you can also use /screen for a more radical compression (i.e. lower quality). Enjoy it!

[Jean Carlo Machado]

convert -density 200 -colorspace GRAY worked like a charm for me