recognize text without losing quality

[KazuT]

In Acrobat DC 2015.17.20053, I created a PDF from a .PNG (resolution 131.6 dpi). It created a clean-looking .PDF (here it is at 150%):

But when I use OCR Text Recognition, it degrades the quality and creates a lot of artifacts:

Is there a way to use Text Recognition and maintain decent quality?

[Test Screen Name]

Which OCR options did you use?

[KazuT · Accepted solution]

It seems that, in order to maintain high quality, I have to choose "Output / Searchable Image (Exact)". The file size stays enormous; but the text doesn't get dirty (too compressed).