I have no idea what your saying. The white pixels match the pixel size in the D. And you are telling me that if we combine a gif with jpg we get this result?
moe don’t know either.
We don’t know when the PDF/OCR effect to this doc happened. Could have been when it was scanned, or at the HI DOH as part of some digitization process.
But when it did, not all of the black on the doc was processed. This happens when the weight (thickness) or density (darkness) is too insignificant.
Those areas not processed remain on the background layer. Their edges remain antialiased (fuzzy).
Those areas that ARE processed are sharpened (similar to a gif) so that the OCR software can discern what corresponding font character to replace it with. However, for some reason this last step was not taken.