PDF conversion which respects images/tables?

[fivebells]

I would like to convert PDF documents to the native Kindle format, preferably using linux tools, but windows tools would OK too. Both the Amazon email-based conversion service and Calibre mangle tables and images pretty badly. I've heard that converting from html often works better. pdftohtml does a beautiful job of converting pdfs to html when I pass it the "-c" option. It generates a large number of html and image files which the Amazon conversion service doesn't recognize as a single document. Calibre can work with the output of pdftohtml if I provide it as a zip file, but the mobi-format files it generates still have mangled tables and no images. Is there something else I should try?

[susan_cassidy]

If it tries to create HTML tables, I don't think Kindle supports them well, or possibly not at all. If you save the table as a graphic, and embed that via an <img> tag in the html, it might help some. I know if you use Mobipocket Creator, you have to specify all the individual files separately as input, so it's kind of a pain if there are a lot of them. You might ask on the Calibre forum for more Calibre-specific questions.

[jocampo]

Hi There,

Converting PDFs, even to ePub, it's a "hit or miss" conversion. It depends more of the document layout than the tool itself.

Using Calibre, these are the best format to convert, in that order: LIT, MOBI, EPUB, HTML, PRC, RTF, PDB, TXT, PDF.

As you can see, PDF is the last one, because the vector images and its own technical specs.

[Blog Kindle]

PDF conversion is a mess, because PDF not meant to be converted and refolown. It was originally intended as electronic paper that contains immutable documents.

I'd just use Kindle native PDF support and not convert things.

[Bjorn2Read]

If you are willing to go beyond free tools - purchase Nitro PDF Professional (around $60 online). It it a robust PDF conversion tool: converts PDFs to Word, rtf or txt with very good results. Also, it can grab images from the PDF file and save them to a folder (great if you want to reference them in an html file).

[fivebells]

Thanks, Bjorn. The promised features are exactly what I need, so I will check out the free trial.

[charonme]

how does the kindle display your original pdfs? Are the tables mangled too or it looks exactly the same as if it was printed?

Originally Posted by susan_cassidy

If it tries to create HTML tables, I don't think Kindle supports them well, or possibly not at all. If you save the table as a graphic, and embed that via an <img> tag in the html, it might help some. I know if you use Mobipocket Creator, you have to specify all the individual files separately as input, so it's kind of a pain if there are a lot of them. You might ask on the Calibre forum for more Calibre-specific questions.

I think MOBI has very limited support for HTML tables. It also has a limited support of CSS. It certainly is not as robust as ePub.

@fivebells, But regardless, tables are going to be problematic on any small screen device in any format. There just isn't enough room to place more than a few columns on the screen. And even with say only three columns, the text has to be really small and/or the text wraps so often as to make the table look ugly.

If possible, I suggest rewriting the tabled information into a non-tabled format.

[GlassX]

Originally Posted by charonme

how does the kindle display your original pdfs? Are the tables mangled too or it looks exactly the same as if it was printed?

It looks exactly the same. However in .pdfs with two columns (e.g. scientific papers) it's either too small to read or too big and needs much scrolling.

[baccilus]

Originally Posted by GlassX

It looks exactly the same. However in .pdfs with two columns (e.g. scientific papers) it's either too small to read or too big and needs much scrolling.

I have observed that two column science papers are easier to read because we can use 200% or more zoom level. However, some full page pdf's can only be read in the landscape mode which I hate.