I was testing Orca's reading in PDF files, and I noticed it doesn't read header levels (like level 1 title, level 2 titles, etc). Is this a problem with Evince, Orca, some Orca configuration I missed, or the API?
I can't read alternative text from images from PDF files too.
I noticed this while in GNOME 3.18, but I have this with GNOME 3.20 too...