PNG to PDF

PNG and JPEG take opposite routes into a PDF. JPEG data can be handed over untouched, but PNG uses its own filtering scheme ahead of Deflate compression, which PDF image objects do not replicate. So a PNG is decoded to raw pixels and then re-embedded using PDF’s own Flate-compressed image object. The important part is that both halves of that trip are lossless: Deflate in, raw pixels, Deflate out, with no quantisation anywhere. Every pixel in the PDF has the exact value it had in your file, which is why PNG remains the right source for screenshots, charts, line drawings and logos — the hard black-on-white edges that JPEG smears into grey halos stay crisp. Two consequences follow. Alpha transparency has no equivalent in a printed page, so transparent regions are composited onto white. And because PDF’s Flate cannot use PNG’s cleverest row filters, the embedded image is often somewhat larger than the source file.