Willus.org Home   |   Archive   |   About  

Willus.com's K2pdfopt Help Page

Return to K2pdfopt Home Page

Getting Started:
    1. Windows
  Text Menu
  (now with video!)
  2. Linux
  Env vars
  3. Mac OSX
  4. Help in other languages

FAQ | Forum

Customizing K2PDFOPT:
    1. K2pdfopt GUIs
    2. Disabling the Windows GUI
    3. The interactive menu
    4. List of command-line options
    5. Using a shortcut
  (now with video!)
  6. Using the K2PDFOPT environment variable
  7. Using the command line

Adjusting the output:
    1. Screen Size
    2. Increasing the magnification
    3. Landscape mode
    4. Output File Size
    5. Setting Margins
    6. Color Output
    7. Uneven Line Breaks/ Excess Margins

Processing Options:
    1. Showing Markings
    2. OCR
    3. Native PDF
  (now with video!)
    4. Auto-Straightening
    5. Ignoring Borders/ Headers/Footers
    6. Detecting Columns
    7. Protecting Regions
    8. Column Order
    9. Right-to-Left Page Scanning
    10. Using Ghostscript

As of v1.60, k2pdfopt has an option so that rather than rendering the output file as a sequence of bitmaps, each output page is rendered directly using the source PDF file instructions, but with translation, scaling, and cropping directives to place the source regions at the appropriate places on the output pages. This is especially beneficial if the source file is a native text PDF file (rather than a scan of a document). In this case, the text in the output file will be searchable, selectable, and have the identical fidelity to the input file. On the other hand, native PDF output is not compatible with text wrapping, contrast adjustment, gamma adjustment, or sharpening. Also, if too many pieces of a single source page are pieced together on one output page, it can be difficult to select or search the text as well, it takes longer to render, and it may even cause your device to run out of memory, hang, or re-start. The video below covers the examples that follow it.

UPDATE: In all of the examples below, with k2pdfopt v2.x, OCR is no longer necessary to get searchable text in the default k2pdfopt output (with re-flowed text). If the source PDF document has searchable or highlightable text (as do all of the examples below), then either k2pdfopt output type (native PDF or the default re-flowed text mode) should also have searchable text without having to resort to time-consuming OCR.


Below are three examples of using k2pdfopt to convert a two-column scientific article. The default conversion (top left) creates rasterized output and has selectable text because I used the -ocr option. The first native output conversion (top right) uses -mode 2col and is the native output option which has the best select/search capabilities (after the first page). Try viewing this PDF at very high magnification. You'll see that it does not lose fidelity, unlike the default conversion. This is one of the benefits of native output. The third conversion, with -vb 1.5, scales some of the text to be more readable, but is much more difficult to select text in. It's also larger and takes longer to render than the other native output option, which defaults to -vb -2. The -vb -2 option preserves all vertical spacing in each source region and therefore requires fewer distinct cropping regions to be placed onto each destination page. With -vb 1.5, on the other hand, k2pdfopt places lines of text individually in the destination file rather than entire regions, which has the benefit of scaling some lines to a more reasonable size and fitting more text onto each page. The -mode 2col command-line option (selectable using "mo" on the user menu) is shorthand for -n -wrap- -col 2 -vb -2 -t.

Default (non-native) with -ocr option. (409 kiB)
Native: -mode 2col (129 kiB)

Native: -mode 2col -vb 1.5 (167 kiB)

Below are two examples for converting a single-column text file. The one on the left again uses the default conversion (rasterized output) with OCR and text wrapping. The conversion on the right uses -mode fw (fit width), which has native output and emulates the same option on the soPdf program, eliminating the left and right margins and turning the file on its side (landscape mode) for enhanced readability. Notice that the native output file is much smaller in this case. The -mode fw command-line option (selectable using "mo" on the user menu) is shorthand for -n -wrap- -col 1 -vb -2 -t -ls.

Default (non-native) with -ocr option. (255 kiB)
Native: -mode fw (48 kiB)

The final example shows a 2 x 2 gridded conversion of a two-column magazine article (excerpt). Again, the conversion on the left is the default conversion (rasterized output) with OCR and text wrapping for comparison. The conversion on the right uses -grid 2x2x4, which grids each page into 2 columns by 2 rows with 4% overlap, quite similar to what can be accomplished with Cut2Col. The advantage of gridding is that only one source region will be placed on each output page, which makes text searching and selection the most reliable. Note in this example, because the source file has high-resolution bitmaps, the native output is actually larger than the k2pdfopt rasterized output. Also, notice that the source coloring is always preserved in the native output (not converted to grayscale to save space).

Default (non-native) with -ocr option. (448 kiB)
Native: -grid 2x2x4 (1187 kiB)


This page last modified
Sunday, 28-May-2017 10:09:24 PDT