4

Opening a scanned PDF book with lot of embedded JPG pictures with original Adobe Acrobat is something slow even if loading poor quality jpg embedded pictures and a SSD 500-1500 MB/s fast.

Generally a single core is used during the loading and scrolling process, and cache isn't kept very updated so browsing and scrolling across multiple pictures results as a pain.

For my work I often need to browse between lot of full graphic (no ocr or txt) pdf files.

What I would like is a light and fast pdf viewer that I could use instead of native Acrobat PDF Reader.

I want would like pdf being loaded instantly like when in some softwares you open a dir of JPG files. Keeping the images in cache, use even a lot of ram and all of my multi core processor is something that I would like.

Even using a bit of the GPU is a nice addon.

Thanks!

Requirements: Windows x64 standalone without crazy ui, I prefer a simple interface like original Acrobat software.

1

Chrome's built-in PDF viewer seems to be multi-threaded. Working with 3 files (133 MB, 105 MB, and 102 MB) it seemed to keep them all in cache so scrolling was no problem. The PDFs I tested with were very image heavy, but I am not sure if they meet your definition of "large."

I would like to be able to suggest an easy way to force chrome to prefetch all of the PDFs in a folder as you mention, but that does not appear to be possible without creating some sort of local web scripting (eg: javascript) and leveraging the prefetch HTML attribute, which should suggest to Chrome to cache everything right away (YMMV, Chrome and your OS will negotiate for RAM and Chrome might not get everything you want it to).

  • I'm after a standalone software, btw I tried chrome with some of the PDFs and the scrolling speed seems the same of original Acrobat working in single thread :( – user3450548 Feb 23 '17 at 13:39
  • @user bummer. how big are the PDFs you're working with? you have plenty of ram so that an application could cache them, right? – Michael Feb 24 '17 at 7:41
  • around 120 MB each yeah the cache is here but is still ridiculus to me the scrolling and rendering speed of it. How to put it.. maybe the cache is filled with the compressed versions but you have still to wait the decoding of the files, the app should imho prepare a low res bitmap version in ram to be used during scrolling and reviewing things in order to speed up everything.. also I have a quad core processor, I expect that the jpg decoding is carried out multicore or if this is not possible that each core decode a page so I have 4 pages decoded at same time, that would be a start.. – user3450548 Feb 25 '17 at 9:30
1
+50

Two softwares that I recomend trying are foxit reader and Evince, because they are lightweight and I used them to open files really huge (500Mb+). Also try Sumatra reader , wich is based on mupdf, a very fast library.

But keep in mind that you can have just a misconfigured viewer; If you really only open heavy PDFs that have been scanned, you can disable the aliasing option in configurations. This will reduce severely the amount of processing to show the page, and on your case, without changing the quality of visualization. Note that some scanned PDFs also have a hidden layer with text obtained through OCR, and in this case you should try to see if the results are what you expect. Also, the fact that your performance is bad on Chrome too seems that you can have a misconfigured plugin, or worse, a misconfigured GPU.

  • For what i just tested Evince is super slow with a "loading" bar. And foxit reader is not speedier than original Acrobat Reader :( There's no OCR layer on them, they are just JPGs. Let's say I use a software like ACD See for a Folder full of 2 MB JPGs, it is super fast. Instead even with a pdf full of 200kb images the reading is slower. As far i know all pdf software out of there sucks on scrolling images, rendering them and use massively the cache to allow you scroll faster :( – user3450548 Feb 28 '17 at 1:50
  • I used to see pdfs really big, and it was not even close to the performance you are experiencing... And it was a Pentium4. Are you sure you dont have a misconfig somewhere? The comparison between pdfs and jpgs are not really valid because pdfs have to paginate, the image viewer don't. It will be faster all the time. – flavio Feb 28 '17 at 9:19
  • If you could put an screenshot of the performance monitor before loading the pdf and one with it loaded it will be usefull to understand your setup too. Also inform your version of windows, and if a workflow modification is possible (other ways to do the same task). – flavio Feb 28 '17 at 9:48
  • Let's make a deal, cause the bounty is ending in 21 min is pointless do not assign it. I will give it to you, but please help me find a solution for it given that is not something that can be closed in this 21 min ^^ I will post you a video showing the rendering speed as soon I have 5 min to deal with ffmpeg and acrobat, ok? – user3450548 Feb 28 '17 at 10:57
  • Ok, but first, do you really HAVE to use pdfs? I mean, if you transform pdfs into something else in batch and then you work on whatever format it is, will that accomodate on your workflow? – flavio Feb 28 '17 at 11:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.