Skip to main content
Participating Frequently
April 14, 2021
Question

Very strange behavior - Acrobat DC and Reader scanned documents

  • April 14, 2021
  • 5 replies
  • 1261 views

I cannot seem to get an answer from Acrobat chat help. Perhaps someone can give me a clue here. When I open any file from the American Chemical Society archives collection -- older articles that have been scanned -- as I scroll I see very oddly placed font text bits in front of the image, as though Acrobat is doing real-time OCR and for some reason showing me what it is detecting. Inner metadata reads Producer(PDFlib 7.0.3 \(.NET/Win32\); modified using iTextSharp 5.0.6 \(c\) 1T3XT BVBA).

 

Document: https://drive.google.com/file/d/1QQJQjvJBHn4U84I_-OuD9D7X7T1S3cz5/view?usp=sharing

My experience (MP4 screen capture) https://drive.google.com/file/d/1QQJQjvJBHn4U84I_-OuD9D7X7T1S3cz5/view?usp=sharing 

 

This makes it very hard to read. Any ideas on what this is or why it is happening or especially how I might turn it off would be appreciated.

This topic has been closed for replies.

5 replies

ls_rbls
Community Expert
Community Expert
April 22, 2021

Is this happening in  a Lenovo computer by any chance, and that is using an AMD processor (not Intel) with Microsoft Windows 10?

JR Boulay
Community Expert
Community Expert
April 22, 2021

In addition, this document is not optimized, you can easily half its size with PDF Optimizer from Acrobat Pro.

Try with this optimized version: https://documentcloud.adobe.com/link/track?uri=urn:aaid:scds:US:c0b01ad4-a977-4f59-a539-8bbdba8b9eb2

Acrobate du PDF, InDesigner et Photoshopographe
Participating Frequently
April 22, 2021

That is helpful. But same problem there. The initial view, in fact, without scrolling, was the font-substitution garble. Static. Unreadable. Then the first scroll took care of that. So, surely, that is an Acrobat bug of some sort. No one wants to see this funky font image remain after the document is opened, I think. The trouble, of course, is how to reproduce that for checking.

 

Am I right, though, that the document itself only has the image stream, not any actual text stream? Or are both present in the document? This issue seems to be specific to scanned documents. If I could figure out what is odd about this particular sort of document, I could go back to the publisher and let them know their documents have this issue.

 

I have a pretty high-performance laptop. But it is true that other software is unusually sluggish with screen writing at 4K resolution.

 

Participating Frequently
April 22, 2021

Attaching two images, illustrating the first-display issue. This was not transient; it only disappeared when the page was scrolled or the zoom level was increased. I have seen this bug on several documents that are scans of older journal articles.

JR Boulay
Community Expert
Community Expert
April 22, 2021

It's not real-time OCR, it's just Acrobat which displays the substitution font first, and then the "real" font in a second step.

 

My computer (Dell Inspiron 2020) must be more powerful than yours because I can't see the first display with the substitution font, even when scrolling through the pages very quickly.

Acrobate du PDF, InDesigner et Photoshopographe
JR Boulay
Community Expert
Community Expert
April 14, 2021

This document works fine for me, I can't reproduce this issue.

Acrobate du PDF, InDesigner et Photoshopographe
Participating Frequently
April 21, 2021
Bernd Alheit
Community Expert
Community Expert
April 14, 2021

I can't see any screen capture.

Participating Frequently
April 21, 2021

I think my GoogleDrive idea for sharing this video did not work. See the YouTube link about 20-30 seconds in you will see me scrolling down the document and the fonts and figures completely transforming. There are instances when this is not transient but instead persists after scrolling. The document is from https://pubs.acs.org/doi/pdf/10.1021/ja00721a023

Participating Frequently
April 21, 2021

to add to the observations: I have a 4K laptop that is showing this behavior. My primary question is simply: "What am I seeing here?" What is Acrobat doing that produces this behavior? Some sort of layering? On-the-fly but not fast enough OCR?