Known Issues

Color Shift

A color shift can sometimes be observed between original PDF files and their corresponding DjVu® file.

Here are two related considerations:

When you view DjVu documents, calibrate your display preferences:

  1. From within the DjVu Browser Plugin window, right-click and select Preference.
  2. Adjust your Display Preferences (top left of preferences window).

Sometimes color differences are inevitable. This is typically the case only with input PDF files that are encoded using the CMYK colorspace. Such PDF files are usually "Prepress Quality" PDF files i.e., PDF files that were created for the purpose of being printed. When viewing such files on a CRT screen (monitor or flat panel), the CMYK color space needs to be converted to RGB. Since neither RGB nor CMYK are calibrated color spaces (in general), this CMYK to RGB conversion can be done very differently by different programs. Adobe Acrobat does one type of conversion to display such PDF files. pdftodjvu and the underlying GhostScript libraries use a slightly different conversion algorithm.

 

Font Substitution Issues

Some fonts do not look good after conversion to DjVu, or look different than in the original PDF file.

This can happen when the GhostScript font-matching algorithm pdftodjvu relies on is unable to find an adequate substitution. In this case, the best option is to use Acrobat or Acrobat Reader to create an intermediate PostScript file, and then convert this PostScript file to DjVu.

If you are a Unix/Linux user of pdftodjvu, this process can be automated using the acroread option of pdftodjvu.

If you are a Windows user of pdftodjvu, some manual interaction is required. You can either

OR

This will give a true rendering of the PDF document, but hyperlinks and bookmark information will be lost.

 

Text from OCR Layer Not Readable

PDF documents that contain text extracted from an OCR layer can not expose this text to the current version of pdftodjvu. No text is extracted. However, the hidden text layer can easily be regenerated in DjVu using djvujoin or djvubundle (standard tools in DocumentExpress Enterprise Edition) with the ocr option.