Q: All pages in the source PDF document seem to have a transparent background but in the converted XOD document, most pages have a transparent background, but some seem to get a white background (they are rasterized).
We need that transparency, because we display two documents on top of each other.
A: In this case pages 22 and 30 are being converted into an image, in this case because the converter deemed there were too many elements (to guarantee fast viewing). To prevent this from happening your best choice is probably using --norender (or SetRenderPages(false)), or alternatively, increase the element limit, using --element_limit 10000 will work for the file in question.