changed font or text in pdf import

Bug #429709 reported by Gordon
34
This bug affects 7 people
Affects Status Importance Assigned to Milestone
Inkscape
Fix Released
High
Unassigned

Bug Description

This relates to Inkscape 0.46 running on Windows XP Professional Service Pack 2.
        I can open the same pdf file in Inkscape and Adobe Acrobat 8.0. But in Inkscape, the embedded images show up with a different font. Also in the main document, Inkscape shows some of the non-alphabetic and non-numeric characters differently.
        Open the attached pdf with each program, and compare the bulleted items in the center column of text. Also compare the dimensioned sketch just above the table, on the right side of the page.

Revision history for this message
Gordon (gferris434) wrote :
tags: added: import-export pdf
Revision history for this message
Guillermo Espertino (Gez) (gespertino-gmail) wrote :

Although this bug was reported against 0.46, the situation didn't improve much in 0.47 pre.
The new feature that chooses a font with a similar name, supposedly if the exact font isn't installed sometimes misses the installed font and changes the imported text to a different font.
I'm attaching a PDF file created with Scribus using the free font Delicious (http://www.josbuivenga.demon.nl/delicious.html)
The PDF file was created in Scribus in the same computer, with the font installed (I checked and it's available in inkscape).

Revision history for this message
Guillermo Espertino (Gez) (gespertino-gmail) wrote :

I found something interesting:
When the font is an OTF font, it completely misses the right family. Font substitution doesn't work.
But when a font is a Truetype it uses the right family, but with a wrong variant.
I created a new document with "testing with truetype" using Frankling normal for "testing with" and Franklin Heavy Italic for "truetype" and the imported text was inverted, the hevy italic variant for the first part and the normal variant for the last word.

I'm concerned about how destructive can be this bug in a document with lots of text (and a wrong substitution could be missed). In my oppinion this is a high importance bug, along #168658

Changed in inkscape:
status: New → Confirmed
importance: Undecided → High
Revision history for this message
AnnaMarburg (marburgster) wrote :

I have encountered what I believe is a similar problem.

This happened to me with inkscape-0.47pre3-2.leopard+.dmg; OSX 10.6.1.

It appears that Inkscape thought that "helvetica CY' was closer to 'HelveticaNeueCE - Roman' than just plain helvetica.

As CY stands for 'cyrillic', this made my english text quite hard to read!

It would seem like there needs to be some sort of logic indicating when do select a variant from the same family and when a different family (but in the same language) would be 'closer' - or, god forbid, just ask the user what they'd like to substitute.

Revision history for this message
Galya (galbinr) wrote :

I have encountered the same problem. After running some tests (on windows xp), I also spotted that OTF fonts are not read when importing pdf files.
Opening a pdf which was created in Acrobat Writer (a pdf writer that uses TTF fonts), is imported just fine to Inkscape 0.47-3. Whereas opening a pdf file which was created from Cutepdf (a pdf writer that uses OTF fonts), resulted in squares. Both files were written in Hebrew from MSword.

Revision history for this message
r.koebler (r.koebler) wrote :

Workaround, if you don't need to edit the text:
Use pdf2svg, and open the resulting svg in inkscape. This converts all text to vector-graphics, so the fonts then look exactly the same as in the PDF. The drawback is that since the text is converted to vectors, you cannot edit the text "as text".

pdf2svg: http://www.cityinthesky.co.uk/opensource/pdf2svg/

Revision history for this message
Jabiertxof (jabiertxof) wrote :

Can r.koebler workarround feature be added as option?

Revision history for this message
su_v (suv-lp) wrote :

On 2015-01-26 22:41 (+0100), Jabiertxof wrote:> Can r.koebler workarround feature be added as option?

Already exists (inkscape's "new" (now visible, optional) poppler-based import in trunk and 0.91 is basically almost the same as using pdf2svg (or pdftocairo -svg) on the command line. The text is outlined (and subset with clones)

Can also be done e.g. via evince (by printing to SVG file).

Revision history for this message
Philipp Burch (phlibi) wrote :

I don't know if this is the same bug, but I have a problem importing PDF or PS files generated by LTspice (http://www.linear.com/designtools/software/): When importing such a file in Inkscape 0.48 or also in 0.91, the content is correctly shown in the preview window, but in the main window, all text boxes only contain kind-of-random characters (look like control codes present in the original PS file). Please find attached a PS file showing this behaviour. If I convert it to PDF first using ps2pdf, the same happens as well. Other programs like evince or the Nautilus previewer show the file correctly. Converting the file to SVG first or using the new 'Import via Poppler' option work nicely, so there is a workaround at least.

Thanks.

Revision history for this message
Philipp Burch (phlibi) wrote :

File from above converted to PDF.

Revision history for this message
Bart Van Audenhove (bart-vanaudenhove) wrote :

Added another problematic PDF import where text (and also an image) are not correctly imported.
See the "CE" sign bottom right which changes to "WE", and the trashcan that disappears.
(Note: images copyrighted and not free for use. Please use with discretion only for bug fixing).
Inkscape 0.91 on Windows 7 Pro 64 bit.

Revision history for this message
Bart Van Audenhove (bart-vanaudenhove) wrote :

SVG import of above PDF.

Revision history for this message
Bart Van Audenhove (bart-vanaudenhove) wrote :

(Second attempt to post PDF...)

Revision history for this message
DeMO (devinmobrien98) wrote :

Is this problem occuring in v0.91?
I was thinking it could be a problem with your pdf. The font used in your pdf may be linked instead of embeded which may result in a change of font style.

Revision history for this message
Raja kedia (rajakedia1) wrote :

This problem is due to absence of the given font family in the computer. If you copy and paste the same text from the pdf in MS word, you will find that whole text is changed in some different symbols and characters. I think to solve this, we have to encode the given text in a font that is availabe in computer.

Revision history for this message
Jonathan Hofinger (jhofinger) wrote :

As the Poppler import exists for some time now, closing this issue as Fix released. If you still experience this issue, please open a report in
http://inkscape.org/report.

Closed by https://gitlab.com/jhofinger

Changed in inkscape:
status: Confirmed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.