6 Replies Latest reply: Jun 19, 2013 1:02 PM by smonrad RSS

    Is Deskewing Not Possible With Searchable Image (Exact)?

    NY2LA Community Member

      I am using Adobe Acrobat X. I am working with PDF's consisting of scanned images, where each page is a single image.

       

      I've noticed that when using Recognize Text, set to Searchable Image (Exact), deskewing does not take place.

       

      Considering that I don't want to use Recognize Text at a setting other than Searchable Image (Exact), it is not possible to deskew from within Adobe Acrobat X? Is it necessary to bring up another application (i.e., Photoshop) via Edit Object?

        • 1. Re: Is Deskewing Not Possible With Searchable Image (Exact)?
          KnightD Community Member

          If you want to deskew using Acrobat you can use the optimize PDF settings, but I dislike it's Deskew method. You can also use a third party plugin like AutoPageX by evermap. It has a manual deskew tool that you can use on a page by page basis.

          The deskewing that the OCR engine does looks best, but if you have it return the exact image it will not deskew. If you are OCRing, do not want to downsample the image, but want to deskew it use the Searchable Image (not exact) mode.

          • 2. Re: Is Deskewing Not Possible With Searchable Image (Exact)?
            NY2LA Community Member

            Thanks, KnightD. Can you tell me more about why you dislike optimize PDF's Deskew method?

            • 3. Re: Is Deskewing Not Possible With Searchable Image (Exact)?
              KnightD Community Member

              I'm using Acrobat 9.5, and when I use the "Optimize PDF" settings (deskew only) on black text documents (scanned images) it returns a deskewed image, but it is garbled up with hundreds of jpeg artificats. I get this tiling of darker blocks of black and lighted blocks of gray. It does not look good or print well.

              • 4. Re: Is Deskewing Not Possible With Searchable Image (Exact)?
                NY2LA Community Member

                I did a few quick tests, using Optimize Scanned PDF. I turned everything off except for the Deskew filter. I unchecked apply adaptive compression. The resulting files were reduced in size, suggesting compression had taken place anyway.

                 

                Preflight revealed that CCITT Group4 encoded (black & white) originals were re-encoded to JPIG2 compression, despite my having unchecked apply adaptive compression.

                 

                Preflight revealed that a DCT-encoded (JPEG) original was re-compressed to JPEG2000 (non-compliant with PDF/A-1b), despite my having unchecked apply adaptive compression.

                 

                I wonder whether Optimize Scanned PDF is intended to be used only during scanning, not on already scanned PDF's.

                • 5. Re: Is Deskewing Not Possible With Searchable Image (Exact)?
                  NY2LA Community Member

                  I did three tests using Optimize Scanned PDF on a PDF that I had created years ago with my Epson scanner, long before I had Adobe Acrobat. The PDF was 12 pages consisting of a single DCT-encoded (JPEG) image per page, at 300ppi, 19.2 MB in size. For each of these test, I used the Deskew filter, sometimes with apply adaptive compression on, and sometimes with it off:

                   

                  1.) I tested using Optimize Scanned PDF on the PDF, with apply adaptive compression on. I set Color/Grayscale to JPEG, and Monochrome to CCITT Group 4, at High Quality. The file size grew from 19.2 MB to 31 MB. I could definitely see artifacts in the resulting PDF, particularly in the type. Preflight showed pages consisting of two JPEG-encoded images per page, plus "Masked by position" CCITT-encoded images, which probably contributed to both the enlarged size, and the artifacts. Preflight also showed the ppi of the images as 150ppi, despite the enlarged size. Re-encoding a JPEG to a JPEG at High Quality resulting in a larger file is consistent with my past experience. Did it deskew? Yes, with mixed results. This definitely doesn't seem to be the right place to apply adaptive compression. Perhaps it would work on a TIFF. I think it is intended to use during the scanning process, before the scan gets saved to a file.

                   

                  2.) I tested using Optimize Scanned PDF on the PDF, with apply adaptive compression off. I kept Color/Grayscale set to JPEG, and Monochrome to CCITT Group 4, at High Quality. Acrobat produced a message: "Acrobat has detected a page that is already optimized. Do you want to run optimization on this page again?", defaulting to a checked box for "Apply to all compressed pages". I unchecked the box, and clicked Cancel. The process brought up the same message for every page, and I clicked Cancel every time. The resulting file was the same as the original, 19.2 MB. Did it deskew, though? No, it definitely did not deskew.

                   

                  3.) I tested using Optimize Scanned PDF on the PDF, with apply adaptive compression off. I kept Color/Grayscale set to JPEG, and Monochrome to CCITT Group 4, at High Quality. This time, when the "Acrobat has detected a page that is already optimized. Do you want to run optimization on this page again?" message came up at the beginning, I kept the box checked, and clicked okay. I did not like what I observed during the process, the checkerboard pattern coming up on the pages. Very strangely, the file size grew to a whopping 72.4 MB. Preflight showed a single DCT-encoded (JPEG) image per page (as the original), at the same 300ppi as the original. The only thing that I can think of to explain the increased file size is the High Quality setting, again considering re-encoding a JPEG to a JPEG at High Quality resulting in a larger file is consistent with my past experience. Did it deskew? No.

                   

                  To me, Optimize Scanned PDF doesn't seem to be intended for use on an already created PDF.

                  • 6. Re: Is Deskewing Not Possible With Searchable Image (Exact)?
                    smonrad

                    Deskewing pages of documents scanned into a pdf file -

                     

                    Using Acrobat X Pro, if you do an Alt/Edit/Find and enter a word in the document in the pop-up Find bar

                    and

                    the scanned document image is too skewed to search,

                    a popup says   the document is not set up for OCR, would you like to set it up for OCR?

                    and

                    if you select Yes, then it will deskew each page and show the progress in a bar in the

                    lower right corner.

                    This works well and straightens each page without any problem.

                     

                     

                    This will not work in a Binder or combined multiple scanned image situation in one file,

                    the search will just process any page(s) in the file which are searchable

                    and

                    ignore the skewed page.  So the skewed page(s) have to be searched and deskewed this way

                    before combining.

                     

                    This is the only deskew operation we have been able to find in Acrobat X Pro.