Pdftotext из xpdf работает нормально для обычного файла встроенных шрифтов, но не работает там, где есть шрифты встроенных подмножеств. Есть ли обходной путь для этой проблемы?
The issue is probably that the characters which are rendered using the subset font have a custom encoding - the numeric representation of the characters does not correspond to ASCII, Latin-1 or any other common encoding.
See
This means there isn't an easy workaround.
In this situation, I have printed the PDFs using the Adobe PDF printer via a high resolution (1200 dpi+), high quality image(up any settings you can). Then, I OCR the image PDF leaving me with a searchable and workable PDF.
When I have many PDFs to do over thousands of pages, I have opened multiple PDF windows at once to do this simultaneously using multiple cores for multiple PDFs. It is a PITA, but it works.
Hopefully your files are small! I've done this to upwards of 10,000 pages once (building code books). Not fun.