For some reason, I've encountered an issue where pdf's scaled using 'custom' page size don't seem to allow post-application of PDFX items, while when I use a standard 11x17 size, it does!?":

$ pdfscale -v -r 11x17 pspread-booklet.pdf

$ gs -sDEVICE=pdfwrite -dPDFX-r1200 -g20400x13200 -sColorConversionStrategy=CMYK -dProcessColorModel=/DeviceCMYK -sOutputFile=gs-converted-pdfT.pdf pspread-booklet.11X17.pdf -c "[/PAGES pdfmark << /PDFXSetBleedBoxToMediaBox false /PDFXTrimBoxToMediaBoxOffset [171 63 171 63] /PDFXBleedBoxToTrimBoxOffset [9 9 9 9] >> setdistillerparams"

^^ inspection of such version ^^ with pdfinfo -f 1 -l 1 -box gs-converted-pdfT.pdf shows boxes have been updated

$ pdfscale -v -r 'custom in 15 10' pspread-booklet.pdf

$ gs -sDEVICE=pdfwrite -dPDFX -r1200 -g18000x12000 -sColorConversionStrategy=CMYK -dProcessColorModel=/DeviceCMYK -sOutputFile=gs-converted-pdfM.pdf pspread-booklet.CUSTOM.pdf -c "[/PAGES pdfmark << /PDFXSetBleedBoxToMediaBox false /PDFXTrimBoxToMediaBoxOffset [27 27 99 99] /PDFXBleedBoxToTrimBoxOffset [9 9 9 9] >> setdistillerparams

^^ inspection of such version ^^ with pdfinfo -f 1 -l 1 -box gs-converted-pdfM.pdf shows boxes are all still equal to MediaBox

Right now when I resize a document that has some full bleed pages, I get white bars on the left and right side. Is there any way to have it fill the page and just cut off the top and bottom of the page when resizing?

Running ./Makefile gives me the following output and then, of course, nothing works.

Dragging / dropping the pdfscale.sh manually into the usr/local/bin directory makes it work but I guess this is not intended.

I am having the following problem:

$ ./pdfScale.sh -s 0.9 infile.pdf outfile.pdf
Unrecoverable error: rangecheck in .putdeviceprops
$ ./pdfScale.sh -v -s 0.9 infile.pdf outfile.pdf
pdfScale.sh v1.0.11 - Verbose execution
Checking dependencies
Scale factor: 0.9
Input file: infile.pdf
Output file: outfile.pdf
Width: 0 postscript-points
Height: 0 postscript-points
Translation X: 0
Translation Y: 0
Unrecoverable error: rangecheck in .putdeviceprops

I have no clue what's causing this or why. The input is a DIN-A4 PDF-File with 105 pages produced by pdfnup.

First of all, thank you for pdfScale, it is really great!

However, I noticed that the repository does not have a licence file. There is a mention in the README file that the software is “free”, but could you specify under what terms the software is released?

Michael J. Cole's code snippet is released under a CC BY-SA 4.0 due to the way StackOverflow works. Thus, correct me if I am wrong, but this project should also be released under a CC BY-SA 4.0. Another alternative would be to release it under the GPLv3 licence, since the CC BY-SA 4.0 is one-way compatible with the GPLv3.

Hi,

I love this tool! Thanks for creating it. 👍

Just noticing that on my macOS, Sierra, I get a failed message for grep:

$ pdfscale -v -v -r 'custom in 5.955 8.515' -f disable -s 1.1 foo.pdf
2017-07-03:23:34:54 | pdfscale v2.1.2 - Verbose Execution
2017-07-03:23:34:54 |    Mixed Tasks: Resize & Scale
2017-07-03:23:34:54 |     Input File: foo.pdf
2017-07-03:23:34:54 |    Output File: foo.CUSTOM.SCALED.pdf
2017-07-03:23:34:54 |  Get Page Size: Adaptive Enabled
2017-07-03:23:34:54 |         Method: Grep
2017-07-03:23:34:54 |                 Failed
2017-07-03:23:34:54 |         Method: Mac Quartz mdls
2017-07-03:23:34:54 |   Source Width: 612 postscript-points
2017-07-03:23:34:54 |  Source Height: 792 postscript-points
2017-07-03:23:34:54 |    Auto Rotate: PageByPage
2017-07-03:23:34:54 |    Flip Detect: Disabled
2017-07-03:23:34:54 |   Run Resizing: CUSTOM ( 429 x 613 ) pts
2017-07-03:23:35:13 |      New Width: 429 postscript-points
2017-07-03:23:35:13 |     New Height: 613 postscript-points
2017-07-03:23:35:13 |   Scale Factor: 1.1
2017-07-03:23:35:13 |  Translation X: -19.499766
2017-07-03:23:35:13 |  Translation Y: -27.863302
2017-07-03:23:35:13 |    Run Scaling: 10 %
2017-07-03:23:35:35 |   Final Status: File created successfully

Is that normal?

Thanks!

The command line argument is made case insensitive by converting it to lowercase but the strings in the script contain uppercase characters in the names of architectural paper sizes so they can never match the argument. I changed the names stored in the script to lowercase and it has been working great. Easy fix. Thank you for the excellent script!

Version of pdfScale:
Gustavo Arnosti Neves - 2016 / 07 / 10
Latest Version - 2020 / 04 / 04

I haven't gotten to the point of making pull requests, so here's my diff output:

1631,1635c1631,1635
< archE 36.0 48.0 914 1219 2592 3456
< archD 24.0 36.0 610 914 1728 2592
< archC 18.0 24.0 457 610 1296 1728
< archB 12.0 18.0 305 457 864 1296
< archA 9.0 12.0 229 305 648 864"
---
> arche 36.0 48.0 914 1219 2592 3456
> archd 24.0 36.0 610 914 1728 2592
> archc 18.0 24.0 457 610 1296 1728
> archd 12.0 18.0 305 457 864 1296
> archa 9.0 12.0 229 305 648 864"
1741c1741
< 11x17 ledger legal letter lettersmall archE archD archC archB archA \
---
> 11x17 ledger legal letter lettersmall arche archd archc archb archa \

Hello,

How do I decrease the execution time of pdfscale

command: pdfscale -r A4 --vert-align center --hor-align center -s 0.92 --image-downsample Bicubic --image-resolution 144 --pdf-settings screen int.pdf out.pdf

Runtime: 0m58708s

Centos7
memory: 252gb
Processor: Intel(R) Xeon(R) CPU E5-2680 v4 @ 2.40GHz

GPL Ghostscript 9.25
pdfscale v2.5.8

Adding more control over the GS call.
Also need to decide on default values (use 600dpi and bicubic?)

Add parameters and default values for:

• DPI ( -dColorImageResolution= )
• Rescaling Method ( -dColorImageDownsampleType= )
• Output Device ( -sDEVICE= )
• GS PDF Settings ( -dPDFSETTINGS= )

Example of dpi/rescaling on @czr's fork: czr@058bfbb

        # Scale page
        "$GSBIN" \
-q -dNOPAUSE -dBATCH -sDEVICE=pdfwrite -dSAFER \
-dCompatibilityLevel="1.5" -dPDFSETTINGS="/printer" \
-dColorImageResolution=600 -dColorImageDownsampleType=/Bicubic \
-dGrayImageResolution=600 -dGrayImageDownsampleType=/Bicubic \
-dColorConversionStrategy=/LeaveColorUnchanged \
-dSubsetFonts=true -dEmbedAllFonts=true \
-dDEVICEWIDTHPOINTS=$PGWIDTH -dDEVICEHEIGHTPOINTS=$PGHEIGHT \
-sOutputFile="$OUTFILEPDF" \
-c "<</BeginPage{$SCALE $SCALE scale $XTRANS $YTRANS translate}>> setpagedevice" \
-f "$INFILEPDF" 

References:

• https://www.ghostscript.com/doc/9.05/Ps2pdf.htm#Options
• https://www.ghostscript.com/doc/9.05/Readme.htm
• https://askubuntu.com/questions/3382/reduce-filesize-of-a-scanned-pdf
• https://gist.github.com/firstdoit/6390547

I am pretty confident that this is not immediately related to pdfScale but rather to a change in one of its dependencies that leads to the outcome.

Expectation:

When a pdf (which in our sample case is a svg file that has been converted to a pdf) gets resized to a new file that has different dimensions than the original one the larger side of the content should fill 100% of the new page while the smaller side should be centered.

Current behaviour:

Resizeing also resizes the content as expected but they do not get centered/aligned correctly.

How to replicate

1. Get a random pdf file (e.g. the Wikipedia Logo (svg > pdf))
   Wikipedia_wordmark.pdf
2. Try to resize the object — e.g. to a object with 700x400pt:
   pdfscale -r "custom pt 700 400" Wikipedia_wordmark.pdf
3. Output
   Wikipedia_wordmark.CUSTOM.pdf

pdfscale v2.5.3 - Verbose Execution
Single Task: Resize PDF Paper
Dry-Run: FALSE
Input File: Wikipedia_wordmark.pdf
Output File: Wikipedia_wordmark.CUSTOM.pdf
Get Page Size: Adaptive Enabled
Method: Grep
Failed
Method: Mac Quartz mdls
Source Width: 375 postscript-points
Source Height: 64 postscript-points
Print Mode: Print ( auto/empty )
Scale Factor: Disabled (resize only)
Fit To Page: Enabled (default)
Auto Rotate: PageByPage
Flip Detect: No change needed
Run Resizing: CUSTOM ( 700 x 400 ) pts
Final Status: File created successfully

OS: macOS Monterey 12.2.1 (21D62)
GS: 9.56.1
Imagemagick: 7.1.0-36 Q16-HDRI arm 20076
BC: bc 1.06

Maybe as a small addon — I have another Mac running Big Sure 11.5.1 with the identical Imagemagick/BC setup but a slightly outdated GS Version (9.54.0) where everything outputs like expected — in case that is helpful.

I've found an odd case where a specific png was being converted into a jpeg when going through img2pdf and pdfScale.sh. I've uploaded it and a test script showcasing the issue over here: https://github.com/RamKromberg/pdfScale.sh_is_lossy

To be clear, I'm not sure if it's even a bug seeing how there's no talk of lossness in the pdfScale.sh docs... And I'm not even clear where the issue lies since both img2pdf and pdfScale.sh are showing some odd behavior with this specific sample... But I figured I'd ask you first since I can still extract a png out of the img2pdf's conversion (albeit, an oddly small one...) but not from the pdfScale.sh's pdf.

Hopefully not wasting your time...

Time to kill all dependencies

We finally have a solution for getting the page size with ghostscript.
Thank you Stefan Dragnev!
https://stackoverflow.com/a/52644056/1273636

PROS

• No need for external dependencies anymore (pdfinfo, identify, etc)
• Should never fail on any PDF
• We have the sizes of ALL pages

CONS

• Seems a bit slower than the other methods

I THINK it is slower also because it is traversing all pages and checking them.
Which means that a version that only gets the size of the first page could be a lot faster.
(need to try and test)
For most operations we only use the size of the first page anyways.

I do want to offer the option for the user to choose the page size though (optional call).

I also want to have the option of listing ALL page sizes (on --info or similar parameter).

Examples calls

$ gs -dNODISPLAY -dQUIET -sFileName=../mixsync\ manual\ v1-2-3.A0.SCALED.pdf -c "FileName (r) file runpdfbegin 1 1 pdfpagecount {pdfgetpage /MediaBox get {=print ( ) print} forall (\n) print} for quit"
0 0 3370 2384
0 0 3370 2384
0 0 3370 2384
0 0 3370 2384
0 0 3370 2384
0 0 3370 2384
0 0 3370 2384
0 0 3370 2384

Without -sFileName

$ gs -q -dNODISPLAY -c "(../mixsync\ manual\ v1-2-3.pdf) (r) file runpdfbegin 1 1 pdfpagecount {pdfgetpage /MediaBox get {=print ( ) print} forall (\n) print} for quit"
0 0 841.89 595.29
0 0 841.89 595.29
0 0 841.89 595.29
0 0 841.89 595.29
0 0 841.89 595.29
0 0 841.89 595.29
0 0 841.89 595.29
0 0 841.89 595.29

Seems like -sFileName is a good idea, since it handles spaces on names.
Also, we may want to include -dBatch and remove quit from the PS script.

Need to test/adapt a bit more and also implement it into the adaptive method.
I will probably use this as second option, if GREP fails (if grep is indeed a lot faster).

I will also probably leave the choice to force the external modes (eg. -m pdfinfo).
It COULD be useful on specific cases.

Help text contains a repeated word "used used"

I checked this nice looking script with shellcheck which spewed a range of warnings (that you might take inspiration from as well, but that's besides this issue) and one error:

In pdfScale.sh line 2045:
        isFloat "$1" && [[ (( $1 > 0 )) ]] && return $TRUE
                                 ^-- SC2071 (error): > is for string comparisons. Use -gt instead.

and indeed, if I feed the value "-1" to that function, it returns true.

Here is a replacement for that function which I believe works properly (without turning to bc as it seems you want to try avoid third-party tools):

isFloat "$1" && [[ "$1" =~ ^0*[1-9] || "$1" =~ ^0*[.]0*[1-9] ]] && return $TRUE

I tried scaling a CalTopo PDF map with geospatial data. The scaling works; I like it.

But the output file was missing the geospatial data.

Using pdfscale to scale a page doesn't preserve hyperlinks.
Could you add the possibility to pass the "-dPrinted=false" parameter to gs?
Now i'm using this script with the addition of this parameter (i edited the script in my local machine) to achieve my goal.
Anyway, nice work! This script saved me hours of life!

Help text contains the word "usefull" which is a misspelling: Should be "useful".

Hi Tavinus,

On macOS (10.14.6) I got an error trying to use the install script.

Fabers-MacBook-Pro-2:~ faber$ curl -s -o /tmp/pdfScale.sh 'https://raw.githubusercontent.com/tavinus/pdfScale/master/pdfScale.sh' && bash /tmp/pdfScale.sh --install
readlink: illegal option -- f
usage: readlink [-n] [file ...]
pdfScale.sh v2.4.9 - Self Install

Current location : 
Target location  : /usr/local/bin/pdfscale
cp: : No such file or directory
There was an error when trying to install pdfScale.
Do you want to try again with sudo (as root)? Y/y to continue > Y
cp: : No such file or directory
There was an error when trying to install the program.

Current location is not detected. This is probably linked to trying to use readlink on macOS.
https://stackoverflow.com/questions/1055671/how-can-i-get-the-behavior-of-gnus-readlink-f-on-a-mac

My workaround to this was by installing
brew install coreutils and then modifying the pdfScale.sh to use greadlink instead of readlink following the answer: https://stackoverflow.com/a/4031502

Best regards,
Fabian

Help text contains the word "Milimeters" which is a misspelling: Should be "Millimeters" (i.e. with double "l").

I'm using your tool inside of a shell script which gets run by an Automator.
I am not the author of this script but it used to work up until 2-3 weeks ago, then now whatever I try the output PDFs are not coming out the size written inside the script.

For example, the one pasted below comes out 611 x 684 instead of the correct amount.
Could you help me find the issue? Thanks!

#!/bin/bash

export PATH=$PATH:/usr/local/bin:$HOME/Desktop
EXT="_Quarto.pdf"
OUTFOLDER="$1"
shift 1

function pdf_rescale () {
    /usr/local/bin/pdfscale.sh -r 'custom pt 657.64 864.57' -s 0.985 "${1}" "${2}"
    return
}

# loop through all input files, or folders and transform A4 PDF to Quarto (232 x 305) PDF scaled to 98.5 percent
for f in "$@"
do
    # is it a directory?
    if [ -d "$f" ]; then
        # nullglob prevents expansion of "*.pdf" to variable x if no match
        shopt -s nullglob
        for x in "${f}"/*.pdf
        do
            shopt -s extglob
            pdf_rescale "${x}" "${OUTFOLDER}/${x//+(*\/|.*)}${EXT}"
            shopt -u extglob
        done
        shopt -u nullglob
    # is it a non-zero, regular file?
    elif [ -s "$f" ]; then
        shopt -s extglob
        pdf_rescale "${f}" "${OUTFOLDER}/${f//+(*\/|.*)}${EXT}"
        shopt -u extglob
    fi
done
exit 0

EDIT: with help from user VikingOSX from discussions.apple.com, I discovered that the issue presents itself only when the input file is a PDF made up combining different PDFs, for example using Adobe Acrobat.
Your tool will correctly resize the MediaBox of the file but will also apply a CropBox to it that maintains the proportion of the original file.
For example: if I want to convert such file from A4 to 9x12in the resulting file available to the user will be 12in high but 8.51in wide, as it will keep the A4 proportions.
Could you please give a look at this issue and tell me how to solve it?
Thank you very much

I have Letter-sized PDFs with content only for 6x9" per page. This content is at the top left portion with white space to the right and bottom. I am trying to scale the content to fill the page. The pages are meant to print small for a booklet but I need a "large print" version at full letter size. Is this possible? I've tried various options with no luck.
I think (not sure) that I need to crop the page to 6x9" then resize to 8.5x11" with a scale of 1.25 to make the contents fill the page.

Tested Systems

• Debian 18 Jessie 32bit & 64bit
• Debian 17 Wheezy 32bit & 64bit
• Ubuntu 16 32bit & 64bit
• Amazon Linux AMI 2016.09 32bit
• MacOS X Yosemite 10.10.5
• Windows 10 64bit - CygWin64
• Windows 10 64bit - Bash on Ubuntu on Windows

Please report your system

It would help me a lot to know in which systems you have successfully run pdfScale.sh.
If you could not run pdfScale.sh in a system, please open a new issue.

Relevant commands

$ cat /etc/*release 2>/dev/null
$ uname -s -r -v -m -p -i -o
$ bc --version
$ gs --version
$ pdfinfo -v 2>dev/null || echo 'No pdf-info'
$ identify --version 2>/dev/null || echo 'No identify'

Template for reporting a system:

Just copy and paste this into your answer and then copy and paste each code block on its place.

# Enter Distro / System name here
#### release
```
$ cat /etc/*release 2>/dev/null
# contents...
```
#### uname
```
$ uname -s -r -v -m -p -i -o
# contents...
```
#### bcmath
```
$ bc --version
# contents...
```
#### ghostscript
```
$ gs --version
# contents...
```
#### pdfinfo
```
$ pdfinfo -v 2>/dev/null || echo 'NO PDFINFO'
# contents...
```
#### identify
```
$ identify --version 2>/dev/null || echo 'NO IDENTIFY'
# contents...
```
#### main use
**Use for resizing:** YES/NO - what papers?
**Use for scaling:** UP/DOWN - printing margins?
**Use for auto-rotation:** YES/NO - which option?
**Use flip-detection:** YES/NO, which option?
**Use adaptive-mode:** YES/NO (why on earth NO?)

Feel free to add/omit stuff, most important thing is to know the distro you are running

Examples below

Variable PDFSCALE_NAME does not properly doublequotes "$0" variable for the subshell running basename command: "$(basename $0)"

When script is run in a folders where some contain spaces or other characters errors such as "basename: extra operand ..." appear due to variable expansion.
The correct way is to doublequote also $0:
PDFSCALE_NAME="$(basename "$0")"

This PR fixes the issue and bumps the version up to 2.5.6: #39

Hi,

I ran a command:

pdfscale -v -v -r A5 -s 0.95 -a none <infile.pdf>

but got the following error that failed to produce output:

2024-05-24:07:30:02 |   Final Status: Error detected. Exit status: 0
PdfScale: ERROR!
Ghostscript Debug Info:
The old, written in PostScript, PDF interpreter has been removed entirely.
You should cease using -dNEWDPF as it has no effect now.
Continuing to process PDF file using the new, written in C, PDF interpreter.The old, written in PostScript, PDF interpreter has been removed entirely.
You should cease using -dNEWDPF as it has no effect now.
Continuing to process PDF file using the new, written in C, PDF interpreter.
   **** Error: Page drawing error occurred.
               Could not draw this page at all, page will be missing in the output.
GPL Ghostscript 10.02.1: Unrecoverable error, exit code 255
GPL Ghostscript 10.02.1: Page object was reserved for an Annotation destination, but no such page was drawn, annotation in output will be invalid.

Thanks and cheers.

To add image content here

I want to scale a PDF and left-align the original content. Is this possible?

Hi,

I have a PDF which was not correctly generated. Pages should be in landscape but they are in portrait mode. All images are crush horizontally and stretch vertically. Is there a way to resize all pages but not rotate them, it's not a problem of rotation, and stretch each page to the new format ?

Thanks in advance.

Wow, your tool is really what I needed for sending an poster to a publisher for fixing the margins. Thanks.

I was wondering if it is also possible/easy to colour the margins created (using scale 0.95) in a specific color, so the resulting image has that colour on all borders for nicer printing/cutting edge?
This is for example useful if the background colour of your poster is not white and to let it flow until the new edge.

Hi,
Not a major issue as things seem to otherwise be working, but I am getting the following error message when running pdfscale to resize some Illustrator PDFs.

PdfScale: ERROR! Ghostscript Debug Info: The old, written in PostScript, PDF interpreter has been removed entirely. You should cease using -dNEWDPF as it has no effect now. Continuing to process PDF file using the new, written in C, PDF interpreter.

The command being run is
pdfscale -v -r 'custom pt 2574 1724' -c full "MBL - trollveggen - A0.pdf" "MBL - trollveggen - 60x90.pdf"

And the verbose output:

pdfscale v2.5.8 - Verbose Execution Single Task: Resize PDF Paper Dry-Run: FALSE Input File: MBL - trollveggen - A0.pdf Output File: MBL - trollveggen - 60x90.pdf Get Page Size: Adaptive Enabled Method: Grep Source Width: 3382 postscript-points Source Height: 2395 postscript-points Print Mode: Print ( auto/empty ) Scale Factor: Disabled (resize only) Fit To Page: Enabled (default) Auto Rotate: PageByPage Flip Detect: No change needed Run Resizing: CUSTOM ( 2574 x 1724 ) pts Cropbox Reset: FULLSIZE ( 2574 x 1724 ) pts Final Status: Error detected. Exit status: 0

pdfscale version is 2.5.8 and ghostscript is 10.02.1 (2023-11-01), running on macOS Sonoma 14.0 and Apple silicon.

I want to resize my pdf to A4 but also i want to check the pages how much extent they are resized. So is it possible to add a border to the original pages in the resized document

Hello,

I had use you awesome script (congrats) to add a right margin to document.
I was able to do it by removing the -dPDFFitPage argument in resize command.

It could be great to add an option to do that in you script.

Best regards,

Manuel

When resizing PDFs, the links are lost.

By the way, thanks for the amazing script!

It would be nice to see the GhostScript command being used. I develop in many ways one being the use of a .Net Wrapper for GhostScript. If I had the command line I could then translate it into the calls for the .Net Wrapper enabling me to match the functionality in my application.

Working with pages

Implement option to "explode" a PDF into several PDFs (one per page).

Same functionality as my pdfSplit.sh script.
https://gist.github.com/tavinus/1a2ae53441b3fa725e93ec043b8b1112

Additionally, give the option for the user to specify a range of pages
on scaling and resizing.

May also be interesting to have an option specific to range splitting,
that just splits the PDF and nothing else.

Options

• --split START-END Range
• --splitAll One PDF per page
• --firstPage / --lastPage Enter start / end on its own (like GS)

These may have relevant info:
https://www.grenadeco.com/tutorial/introduction-to-ghostscript/
https://stackoverflow.com/questions/4826485/ghostscript-pdf-total-pages/4829240

Hi, tavinus. I'd like to package your very helpful script for the Tiny Core Linux repository. Can you please clarify its license? The script just says that it is "free" but does not specify a license.