Keep getting "renderable text" error when I need to OCR PDF's from FrameMaker.

Question

My solution has been to individually extract all those pages, then open them up in Photoshop, flatten them and
widen the canvas size to standard 8.5 x 11.
But that's a little tedious and time-consuming and you have to delete the original page from your document, after
importing the OCR-friendly page.
Is there a printer definition, or something you can set up when you're generating your PDF's in the first place,
that will get rid of that annoying "renderable text" error?

Answer 1

Ok...
I don’t know how it happens but after I save my work in FrameMaker or MS Word, and print
to PDF for the final output, there are often pages with text in them that isn’t recognizable,
or that can’t be found with a CTRL+F search.
That is a serious issue, and one we might be able to help you with, but really, quite separate from the issue here. It's too late to try and fix this once it is a PDF.
What is it that’s lost when OCR is run?
Quality. Small file size. Tags (which might be required legally). Almost everything except the basic text, and that might also be lost given that OCR is not guaranteed to work. This is NOT the right way to solve your problem.
The translator doesn’t have any Adobe products except Reader, so I’m limited to Acrobat
to show her how the words and pictures are laid out on a page.
In order for her to copy and paste that text – or search it, to find all of the places where the
same word might be used – I need to make sure every word is there for her to grab.
I have heard of translators trying to work with PDFs, and few that succeed. You can reasonably expect a transation service to support FrameMaker. But if they don't I recommend you extract the text from FrameMaker to a simple Word or text file. They should be fine using the PDF as a visual reference, and having the text to translate, and for you to flow back into the original layout. (Again, something I'd expect a full service translation to do themselves, but there are advantages to keeping control too).
Those were the 2 pages that gave the ‘renderable text’ error. Don’t ask me why or how, they
look like all the other pages in that document. Except Acrobat thinks they’re scanned graphics,
that’s how they present when you wave the cursor around in them, hunting for text.
Renderable text is just text. It means that somewhere on that page there is text. Surely there is layout, page numbers, whatever from FrameMaker on the pages. If not, we really need to look at your production methods - back to the first point.