Calibre .6 does a much better job of converting pdf than previous releases. PDF will likely need some cleaning up after conversion though, and for this reason epub is probably the most flexible format, since it's just a zipped container with html files. Unzip, edit, re-zip.
The biggest problem with pdf is that many books have page headers and footers. Best way to get rid of this is cropping, but most software only changes the page boundaries, header and footer text is still there, and will show up in the converted file. Not sure whether Calibre's pdfmanipulate function or papercrop crops the hidden text away. The only way I've found so far to remove the cropped text is to use Acrobat Pro's pdf optimize feature, which can remove the hidden text.
I've also had some luck using Acrobat Pro to convert the PDF to HTML. This depends on the quality of the PDF whether it will work though - a lot of PDFs will fail to convert. If it does work then it does a better job than Calibre's built in pdftohtml. Image extraction in particular works quite well. Then you can convert the zipped html to epub or whatever.
|