[pointertovoid]

Hello you all!

Some Pdf documents are seriously old, like typewritten in 1950 and obviously digitized as an image: characters badly aligned, hand annotations, spots...
Nevertheless, Pdf readers like Foxit can find a searched text in them, and so does Google for instance.

So what? Are Pdf readers smart and fast enough to perform the Optical Character Recognition in real time over the whole image? Typical OCR software is slower and less efficient in this process.
Or do Pdf documents contain a text version of the image, performed once when the Pdf is created, and whose position is well-adjusted with the image? Would save time.
Other explanations?

Thanks!