Zotero built-in PDF viewer could not display texts correctly

Lablace · September 16, 2022

I recently added a new book into Zotero and when I tried to view it in built-in viewer, the texts could not be seen. I am sure that they were displayed as I could select and copy the texts, but they could not be seen. That's to say, the text displayed has the same color (white) as the background. What is more strange is some pages could be displayed correctly. These pages each have a graph on it. And the first and last some pages could be displayed correctly.
When I open just this file in macOS Preview and Safari, they are displayed correctly, say, with black texts on white background.
This may be a problem with Firefox as I could not open with newest Firefox on Mac either.
Any solution to this? Thanks very much!

The PDF version is 1.6 if it may help.

Lablace · September 17, 2022

This seems to be a problem with Mozilla/pdf.js as Zotero uses it. Debug console said it has unrecognized color space which I do not quite understand. Whatever, after I opened it in a PDF editor, made a trivial change and undo it, then saved it, everything now goes on the right way. Zotero and Firefox can display the content correctly.

CreativeRaven · October 31, 2022

Having the same issue with Zotero on Mac. Just filed but report 58573826. White on white, but when I highlight the text, it shows up in the annotations.

dstillman · October 31, 2022

@ArizonaNorse: What do you mean by "with Safari on Mac"? This thread is about the built-in PDF reader in Zotero.

If you're having trouble with a particular PDF, please email it to support@zotero.org with a link to this thread.

CreativeRaven · October 31, 2022

My typo, should have said Zotero on Mac. Corrected. More than one PDF.

dstillman · October 31, 2022

We would still need to see an affected PDF.

CreativeRaven · October 31, 2022

sending one example to support address

martynas_b · November 1, 2022

It's not the first time we encounter this issue. The files are always produced with "HathiTrust Image Server" and has some non-standard headers in scanned images. We already reported this issue to PDF.js (our underlying PDF reader library).

CreativeRaven · November 1, 2022

Thank you martynas_b for your quick response and I do hope this can be resolved soon!

bi2022 · March 29, 2023

This issue still persists.
However the pdf.js mentioned in this thread should be fixed by now as my firefox displays the pdfs correctly.

$ zotero --version
Zotero Zotero 6.0.23, Copyright (c) 2006-2018 Contributors

os: ubuntu 22.10

bi2022 · March 29, 2023

A temporary solution for linux users was found in https://stackoverflow.com/questions/3450616/convert-pdf-from-version-1-1-to-1-4-or-higher#58108270

gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -o output.pdf input.pdf

abhinavtripa · January 10, 2024

Same issue here. PDF version of this paper is missing a few characters that are being displayed as expected in Adobe Reader
https://onlinelibrary.wiley.com/doi/10.1002/(SICI)1097-4563(199901)16:1%3C1::AID-ROB1%3E3.0.CO;2-Q

martynas_b · January 10, 2024

@abhinavtripa Try using the Zotero 7 beta if you haven't already, to see if the issue persists.