convert PNG bilevel images with jbig2decode rather than flatedecode
Bilevel images are widely used for scanned ebooks. Plz do so to explicitly reduce sizes.
I experimented with choosing the right encoder for bilevel images at around the time when I started writing img2pdf as is evident from the age of the file
But my tests didn't show a great advantage of using jbig2decode over first using png's paeth filter and then a flatedecode.
But maybe the kind of images I had is wrong -- do you have images of a different type where jbig2decode indeed produces significantly smaller images?
Hi @wizardforcel, I just wanted to ping you about this issue because I'm making a new release soon and was wondering whether you indeed have some input where jbig2decode provides a significant advantage over the current implementation. Otherwise, please close this issue. Thanks!