Home / Major performance issue

I’m using current version of Sumatra. If I try to open PDF document with OCR layers under image, rendered by FineReader then Sumatra takes forever to open such document. Sometimes running optimization over such PDF (i.e. using shrinkpdf.com) might help, but in 90% cases it doesn’t. At the same time Adobe Acrobat Reader and most other readers can open such documents instantly.
So, this appears to be a major performance issue. Hopefully it can be fixed shortly.

P.S.: link to sample PDFs (image only and OCR versions of the same pages) https://github.com/sumatrapdfreader/sumatrapdf/files/1351938/Sample.zip

When commenting on performance issues it helps to provide a link to a sample
I tried to download several pdf samples from the Abbyy site but strangely NONE were produced by Abbyy products
Their User guide was generated by WPCubed
Admin Guide by Microsoft Word
There best User testimonials generated by Adobe products
Hmmmmm

Sorry for missing the sample. Here is a .zip with 2 versions of the same pages: first without OCR layer (only images) and the seconds one is with OCR layer. The last one rendering lags terribly on my machine.
https://github.com/sumatrapdfreader/sumatrapdf/files/1351938/Sample.zip

neremin
thanks for a sample, I agree those files show a problem in time taken to render, however the delay is understandable, I suggest you open a “new issue” on https://github.com/sumatrapdfreader/sumatrapdf/issues

my observations are that the problems stem from the file size since the original images are
(rounded values)
6 MB with pages 20 x 29 cm
whereas the OCR version is only
3 MB with pages 84 x 121 cm

so firstly the 11 kb OCR version has additional images that are 16 times bigger but must be heavily compressed to result in 50% file size, in my experience that is 32 fold compression which considering the structure of the poor tonal background is going to result in phenomenal time wasting for decoding excess baggage

Thank you for your advice. This issue was created with #904

[Why do neremin’s posts linking to the sample archive say “This post was flagged by the community and is temporarily hidden”? I didn’t find anything wrong with the PDFs in the ZIP file. Is this some sort of automated anti-spam thing?]

The irritating slow page rendering issue is seen with many files, not just OCRed ones. Not even sure if Sumatra’s devs can do anything about it since it might well be a MuPDF issue.

Yes, it was automated rule (new user posting multiple links to the same domain) applied by the forum software.

I undid that (I think).