Here are a couple screenshots to demonstrate the process that is causing the bloat.  You can see that Text Recognition picked up stray marks on the edge of the scanned photocopy paper:

I started over with a copy of the PDF version from before using Text Recognition. I did Text Recognition; once again, that increased the file size by only a small amount. Then I started marking these Suspects as Not Text again. Although I didn't do all 216 pages again, I could see the file significantly growing in size after every Save.

Also, whenever I would save, I would see a message "Replacing Fonts", which doesn't make sense. I want to delete these Suspects, as if they were never recognized by Text Recognition in the first place.

LoriAUC wrote:

 

Have you tried just doing a simple Save As > Reduced File Size?

I went back to the 22.9MB file to try this. Indeed, it reduced the file size to 3.5MB, and got rid of most of the bloat.

However, it think it also reduced the quality of the PDF because it added another re-encoding and/or down sampling step, which I don't want to do.

I'd really like to know what designating Suspects as Not Text is supposed to do, and possibly avoid using it if the purpose of it is different from what I want it to do.

I think that I can find it on a Time Machine backup. And yes, if I find it, I will gladly share it.

Indeed, it says maximum file size, 2 MB. The file with the problem is 22.9 MB.

I've sent it to apangasa via www.yousendit.com , because I already have an account.

Thank you, apangasa.

apangasa wrote:

 

Hi,

 

On incrementally saving the file, file size grows and this is in accordance with how feature is designed.

I tried saving the file again with a different name and this time, I could not reproduce the issue. Could you please confirm if you are also seeing the same behavior for non-incremental file save.

 

Thanks

I am not sure that I understand your question. Do you mean Save Versus Save As?

My final step is usually Save As...PDF/A. However, the bloated file was probably from before doing that final step. So I just tried opening the bloated file, and doing Save As...PDF/A, which does produce a file of significantly reduced size. However, I get the spinning beach ball, and have to force quit Acrobat. (I tried it twice.)

apangasa and Test Screen Name, thank you for bringing the difference between Save and Save As… to my attention. This may be my first experience with a program where Save As… means something other than simply making a copy, which can be given a different name and location.

However, now I am wondering whether something else altogether explains the bloating phenomenon.

The majority of the elements found by OCR Suspects are actually correct. As I've been going through all Suspects one at a time, whenever no correction is required, I've been clicking "Find Next", rather than "Accept and Find". Could this be triggering the "Replacing Font" phenomenon, and causing the bloating, when I Save?

Here's an example of an OCR Suspect that is correct, where I clicked Find Next: