Printing to "vector" PDF

[caspertone2003]

Hi everyone.
From time to time I need to get a file with some pages of a PDF I am reading with SPR (PDF are unprotected). Typically I just printed using a PDF printer the pages that I needed. Since some time ago, I noticed that, even if PDF being read was fully vectorial (so I had all the fonts or they were substitued), the newly printed PDF is just made of images (so, you cannot search, or copy, fragments of text).
While I can resource to use other tools so as to split pages or to edit the PDF, this is not so comfortable.
At the beginning I though this was caused by the Microsoft printer to PDF, but most of them I tried do same (Bullzip, PDFCreator, etc.): Intrigued, I found that something similar happens with most browsers. I found a recipe for firefox that if you forbid websites to use their fonts of choice, you are able to print searcheable non image vectorial PDFs.... It looks that Chrome and Edge do not provide a way to escape just printing big and unsearchable pdfs (unless you use a browser with save to PDF feature and you have luck...)

So, while I was tempted to blame W11, MS PDF printer, etc, it looks that at the end it is a behaviour of the applications themselves.

So, my question, is there any way to configure SumatraPDFReader so to output vectorial stream for the PDF printers to create a true PDF non-image output?

TIA

[GitHubRulesOK]

Not really, the old XP vector printing before MS Print to PDF was flaky but if you want to try that, it was last included in version 3.0.

SumatraPDF is built on MuPDF rather than its printing sister app GhostScript.
Thus historically the only way to print Pdf would be using GhostScript which I think has similar issues with WinPrint via MS Print as PDF since it is designed to PDF write vectors without printing OR send PDF vectors to vector printers.

The other long time Print format development team is Adobe with specialist drivers to manage PDF print to other devices.

[caspertone2003]

Thanks for the information, very valuable.
I had a sad face not for the help but for the outcome...
Now I know that SumatraPDFReader does not solve my sections printing needs, so will need to do it with either
a) A GhostScript based tool for printing (will search if available)
b) Will have to use a pdf cutter (such as PDFSam) or a true PDF editor (XChange ones will do)
Such a pity that SumatraPDFReader can not (or decided not to) do it!

[caspertone2003]

Did not find any pdf viewer able to do what I wished. Okular was promising but it also rasterizes printout in Windows...

And the winner is ..... a cutter ... portable, with view of thumbnails, free, opensource....
PDF Arranger - https://github.com/pdfarranger/pdfarranger/releases

[GitHubRulesOK]

that is excessive overkill it is a python wrapper around pykepdf which is a wrapper around qpdf (very good single executable) but that is only a pdf rewriter like GhostScript however all those slower wrappers will have different settings and since I dont know why your altering a PDF dont know if even simpler apps like cpdf or pdfcpu would be faster.

SumatraPDF could easily call cpdf to alter a PDF while viewing or better backup the viewed file and reload an altered one.

[caspertone2003]

Thanks!
I am spending quite long time to find a solution, if Sumatra could do it would be very nice.

My use case and needs are very simple: as an example, let say I have a 90 pages document of which I usually work only with the page number 4 and the page number 64 (number vary depending on document). I wish to create a PDF with pages 4 and 64, so that i do not need to move up and down. I need this new PDF to be searchable, not bitmap. I used to do that opening the PDF in Sumatra, looking the pages I need, then printing those pages to a new PDF... but now I only get bitmaps...
So, I need to have a GUI to visualize the PDF and see what pages I do need, and then to extract couple of pages.
I need a portable tool to do the extraction.

How can I make SumatraPDF to call cpdf ?

Note that i am just a "user" not fully understanding all complexities behind the engines, wrappers, front ends, or whatever. Following your first commend I found SumatraPDF to be build is something called muPDF, not Gostcript, being that the "culprit" of being all prints from sumatraPDF just bitmats. They I found popplar, qpdf, now cpdf, pdfxx, etc... really lost...

Another approach I tried was to use a PDF editor such as XChange one. Just copy the pages I need from the original pdf to a new pdf, save, done. But an editor is much complex ...and very disturbing tool, I am in favor of simple minimalistic user interfaces...

I am unable to understand what you said "SumatraPDF could easily call cpdf to alter a PDF while viewing or better backup the viewed file and reload an altered one.". Do you mean that SumatraPDF code could be easyly be modified to output PDF sections not as bitmaps?

[GitHubRulesOK]

OK SumatraPDF can call external viewer with filename and page number. Lets break down your requirement.

SumatraPDF is a PDF page viewer you use to read pages and decide you want a new file with current page. Say 6

• That is simple we can export the PDF name and current page number 6 to a list
• You also decide later the same new file could include page 64 so we also send that to the list.
• You reach end of file and decide it is just those 1, 2 or say a dozen listed to be merged into a new file so start the list merging

Previously have used similar to convert PDF pages to images

Ok that is 2 distinct actions and there are several methods

• one is collection of numbers (possibly with more than one source file name)
  then blending copies of numbered pages into a new filename.pdf
• two an alternative is split out the single pages as duplicates as required into a folder
  then later blend the desired pages into one or more PDF

There are as usual advantages and disadvantages with both of those 2 but generally quicker than other methods.

The problem with second method is the overheads in resources may create larger results than the first method might. But the advantage over a list is, you visually can instantly see those you may not want or are repeats.

So we go with a listing (that actually would suite GhostScript perfectly as it can merge pages via a text list but may need edits in the listing) and others too can be made to work likewise.

SumatraPDF can Echo the name/numbers as text for example you use an icon or key to call a cmd file or echo direct. along the lines of

that is the easy part and similar to sending just unwanted filenames to be deleted later on exiting SumatraPDF file locking.

The more complex part of sending the list to one of your choice applications is more custom and needs to consider the order of commands and pages both of which are more to your liking?
Things that need to be considered are default page sizes, folder name etc.

[GitHubRulesOK]

Since I suggested GhostScript as an option I saved this discussion as 5 pages and listed pages 2 and 4 to check how well GS performed.
This is just a first example as others may be as good or easier or better output?

Here is the result as 2 pages with hyperlinks and selectable text kept, although it correctly shows errors with the PDF it was given !

Here we can monitor the errors during production, which can be paused to check what problems were encountered

What was needed was a command file to call the latest copy of GhostScript 10.05, I called mine mergelist.cmd

set "gs=C:\Users\K\Desktop\Apps\pdf\GS\gs10050\bin\gswin32c.exe"
"%gs%" -sDEVICE=pdfwrite -o"c:\test\merged-GS.pdf" @"C:\Users\K\AppData\Local\SumatraPDF\plus\GSpagelist.txt"
pause

And the GSpagelist.txt generated by SumatraPDF was

-dFirstPage=2 -dLastPage=2 "C:\Users\K\Desktop\Apps\pdf\SumatraPDF\#4853 Printing to vector PDF.pdf" 
 -dFirstPage=4 -dLastPage=4 "C:\Users\K\Desktop\Apps\pdf\SumatraPDF\#4853 Printing to vector PDF.pdf" 

Which in turn was added to Advanced settings as

ExternalViewers [
	[
		CommandLine = c:\windows\system32\cmd.exe /r echo  -dFirstPage=%p -dLastPage=%p "%1" >>"C:\Users\K\AppData\Local\SumatraPDF\plus\pagelist.txt"
		Name = &Save PDF page # in list
		Filter = *.pdf
		Key = #
	]
]

currently there are no extras like "delete old list"s etc. since it is just to show what is possible

[GitHubRulesOK]

So far I tested Mutool Merge but for this task it is more trouble to script and the text selection is a bit above the line plus not all hyperlinks are kept and a bigger file. So currently consider that as lower value in terms of this solution

Also tried cpdf which had some same issues as MuPDF but hyperlinks are retained but again the file size is bigger still ? (odd perhaps there is something more needed like compress?)

[caspertone2003]

@GitHubRulesOK

(sorry for the bold typography, I was unble to find out why there are there, I cannot remove it...)

Thank you for your time to answer me, you certainly took a lot of effort to help me. As one good turn deserves another I felt enticed to provide this feedback - please do not see it as negative criticism but otherwise!

I will start saying that you have been brilliant in finding a way to extend SumatraPDF very good but very focused/constrained functionality by the use of the external viewers functionality. I am afraid you are "fighting" a very energy consuming "battle" to achieve that end. Certainly I feel that SumatraPDF is not conductive of any deviation to be "just" a pure (although very good one) viewer.

I had to investigate a lot to understand your information (no compliant). You know, none of the versions I tried from SumatraPDF show the "Save as PDF" in the File dialogue, so I was in search of how to have that there. I have now figured that this is the result of using the external viewers way, and I had to understand before that I had to set up your external viewer thing before.

To understand your reply I had to read the answer several times, then I googled a bit and found https://github.com/GitHubRulesOK/SumatraPDF-Plus then I felt I needed to understand your plug in and see how to fit your answer with the plugin thing. Some references for anyone reading this in the future:

• https://github.com/GitHubRulesOK/MyNotes/tree/master/AppNotes/SumatraPDF/Addins
• https://github.com/GitHubRulesOK/SumatraPDF-Plus
• Export PDF as Images with Custom Page Selection in SumatraPDF #4616
• SumatraPDF sending print output to /dev/null #4611 (comment)

(I am sure there are many more along the forum(s) but with those one can manage)

So, to test your approach I did

=======================================
PREPARE THE SOLUTION

STEP1

Set up SumatraPDF to be able to create the list. This involves going to settings/advance options and replacing the standard

ExternalViewers [
]

by

ExternalViewers [
[
CommandLine = c:\windows\system32\cmd.exe /r echo -dFirstPage=%p -dLastPage=%p "%1" >>"C:\userpathforpagelist\pagelist.txt"
Name = &Save PDF page # in list
Filter = *.pdf
Key = #
]
]

NOTE: user has to adjust above to put their userpath where they wish to have their pagelist.txt.
NOTE2: this might break the portability of sumatrapdf portable... unless a scratch C: place is used, then it is up to user need to properly clear the scratch contents.

Step TWO

If one is going to use Ghostcript (there are many other alternatives...)
a) download and prepare https://ghostscript.com/releases/gsdnld.html
b) Create your CMD to join and print the saved pages.
Something along the lines: (save it as .cmd, such as getmypdf.cmd)

set "gs=C:\yourGoshtcryptPATHhere\ghostscriptexecutablenamehere.exe"
"%gs%" -sDEVICE=pdfwrite -o"c:\yournewpdfPATH\younnewPDFname.pdf" @"C:\userpathforpagelist\pagelist.txt"
pause

(remember to substitute paths and names to suit your needs)

========================================================================================
TO USE THE SOLUTION
Now, execution
While seeing your document in SumatraPDF, add the page you want to have to "pagelist.txt", using "File/Save PDF page # in list"
When you are done having all the pages you wish in you "pagelist.txt", execute your CMD, (as above, getmypdf.cmd )
When you are ready, do not forget to delete the now not needed "pagelist.txt"

Yes, it is usable and works. But cumbersome to set up (I doubt the typical user will be able to set it up) and not user friendly/fluid experience to use. LATER ADD: found a way to improve user friendliness (read below in another comment).

I can understand that you see PDF Arranger as an overkill tool to do such simple task as I need. But, from a user perspective

• it is unbelievable that one cannot get a decent printout from any pdf viewer in windows (it is not your fault, and from my view it is a design decision by the SumatraPDF creator(s) that using only muPDF they are killing their viewer future - I will open a request where it should be done.
• following your alternative solution it is much more complex to set up and use that using PDF Arranger or by the use of a PDF editor directly (needing to use a PDF editor for this is quite far away from a Flat design UI point of view)

I am not intending to be argumentative on the degree of "kilness" of each approach, just providing my own sensations for you. Yes, SumatraPDF can be used to generate inputs for other tools to then get the final output. There are many toolkits and cmd sw that allow you to manipulate PDFs to slice and dice a file... But this is not an easy, direct way for a normal user just to generate a new pdf from a given pdf without transforming the contents to a only bitmap based pdf....

(OLD: Certainly I will keep using PDF Arrange, read further below comment)
and I will propose to "enhance" natively SumatraPDF to support vector output to virtual printers.

I wish to thank you a lot for what I have learned following the talk with you.
See you!
CT

[caspertone2003]

Proposal done ...

#4864

[caspertone2003]

@GitHubRulesOK
@ all future readers

Finally, I will use SumatraPDF + GitHubRulesOK solution! (that it is executing arbitrary windows commands using the external viewer functionality). Certainly not easy to set up but quite acceptable and fluid to be used after all.

I set up three commands using the "external viewers" trick.

Prepare: add to settings/advanced options (for reference, if needed, read https://www.sumatrapdfreader.org/docs/Customize-external-viewers )
From next line->
ExternalViewers [
[
CommandLine = c:\windows\system32\cmd.exe /r echo -dFirstPage=%p -dLastPage=%p "%1" >>"C:\YourPathForPagelists\pagelist.txt"
Name = &Save PDF page # to Pagelist
Filter = *.pdf
]
[
CommandLine = c:\windows\system32\cmd.exe /r DEL /Q "C:\YourPathForPagelists\pagelist.txt"
Name = &Delete Pagelist
Filter = *.pdf
]
[
CommandLine = "D:\YourPATHforGhostscript\gswin64.exe -sDEVICE=pdfwrite -o"%d\outputfile.pdf" @"C:\YourPathForPagelists\pagelist.txt"
Name = &Create output PDF from saved Pagelist using Ghostscript
Filter = *.pdf
]
]
<-Up to previous line

Use:
For each page you are viewing that wish to retain: File/Save PDF page # to Pagelist
Once finished: File/Create output PDF from saved Pagelist using Ghostscript
Once happy with the output file: File/Delete Pagelist

When ordered to output PDF, a cmd window with ghostscript giberish will pop up (did not find a way to send errors/warnings to a file), read if all ok or errors or warnings. The you will have to kill. If no errors a "outputfile.pdf" will be created in same folder as your original PDF
If you forget to delete pagelist you will be having unwanted pages in later output files. If there is an outputfile.pdf you might have errors or over writing of an existing file. Beware, it works but it not fool proof ...

Enjoy!
CT
(and thanks againt to @GitHubRulesOK)

[GitHubRulesOK]

To reduce GS "gibberish" which is there to show Console IN PROGRESS problems
one small adjustment is add -q after .exe so ...c.exe -q -sDEVICE=.... removes the version lines which are not in error

There are other simple means to reduce messages but until a method is proven over time that is cosmetics.

[caspertone2003]

My tests with GS go very well, but my pdfs are simple ones. Did not check about size nor hyperlinks. Vector stays vector.
Anyone wishing to try with other tools .... such as qpdf (how to do page selection = https://qpdf.readthedocs.io/en/stable/cli.html#pdf-transformation) or PDFtk Server (how to catenate = https://www.pdflabs.com/docs/pdftk-man-page/), or cpdf (merging and splitting - https://www.coherentpdf.com/cpdfmanual/cpdfmanualch2.html), PDFSam/Sedja (merge&Split - https://sejda.org/ ), or poppler ( https://superuser.com/questions/1321515/splitting-pdf-files-using-pdfseparate-poppler/1000 ), ...

[caspertone2003]

Notice to seafarers

sumatra external viewers interface seems to have its character ... beware!

---

I believe cmd.exe can be used instead of c:\windows\system32\cmd.exe

Did not find a reference for /r switch in cmd.exe /r (seems to do same as /c)
(in my tests /K did not work either)

---

Beware of this:

[
CommandLine = c:\windows\system32\cmd.exe /r DEL /Q "C:\YourPathForPagelists\pagelist.txt"
Name = &Delete Pagelist
Filter = *.pdf
]

In my case, that one deletes BOTH the pagelist.txt and the original PDF !!!
(was unable to find why, I just stopped using it)
As alternative idea, create a generic .bat file that deletes pagelist.txt and it is called from Sumatra with

To generate the new PDF, now I use

[
CommandLine = "D:\YourPATHforGhostscript\gswin64c.exe -sDEVICE=pdfwrite -o"%d\ClonedPages.pdf" @"C:\YourPathForPagelists\pagelist.txt" > "%d\ClonedPages_report.txt"
Name = &Create output PDF from saved Pagelist using Ghostscript
Filter = *.pdf
]

Using gswin64c produces a CMD window report as well as sends to the file ClonedPages_report.txt some kind of log.