Conversion is always difficult because there are so many variables. How was the original created? Is the PDF straight text or scanned images? Remember, PDF is a container and there can be almost anything inside.
If you have a .docx, use Word or OpenOffice to save it as an HTML file, edit the HTML to get as clean a source as possible, then use Calibre or Sigil to convert the HTML to ePub. You may still need to edit the ePub to tweak things.
|