Turning the pdf into PT files is unquestionably the better route. If you were to turn the PT files into pdfs and then compare the pdfs, you’d have to make sure every single line and page broke at the same point–an almost impossible task.
Depending on what language you’re using and how the pdf was created, it’s possible that you could simply highlight the text in the pdf and copy-paste into a text editor. There are also programs which will automatically extract all the text in a pdf. If that doesn’t work you’d have to look into Optical Character Recognition (OCR), where the computer tries to ‘read’ the pdf. But that’s much, much more difficult.
Once you’ve extracted the text from the pdf, you can use a lot of text programs to compare the PT and exported-pdf files. You may need help from a scripter/programmer to remove some of the usfm markers from the PT files so that you can compare the two more easily, but that shouldn’t be too difficult.
