Why are photo PDFs so large?
A modern phone photo is 3–6 MB. Twenty of them make a 100 MB PDF, and nothing about the document needs that: a page printed at A4 and 200 DPI is about 1,700 pixels wide, so anything beyond roughly 2,000 px is invisible detail costing megabytes.
Resizing to a page-appropriate width is the single biggest saving, and it is why the option is on by default.
Will the pages come out sideways?
Phones store orientation as an EXIF flag rather than rotating the pixels. Tools that ignore the flag produce sideways pages — the classic broken photo-to-PDF result.
The flag is read here and applied when the page is built, so a photo taken in portrait becomes a portrait page.
Is this good enough for an official document?
For most portals, yes, provided the photo is legible: flat paper, even light, no shadow across the text, and the whole sheet in frame. Many portals also set a size limit, which the compression option usually clears on its own.
Where a portal demands a scan, use a scanning app or a flatbed — a photograph at an angle is the usual reason a submission comes back.
How do you get the pictures back out?
PDF to JPG renders every page as an image at up to 600 DPI, which is right for printing or archiving a page as a picture.
If what you want is the original photographs rather than rendered pages, extracting the embedded images keeps them exactly as they went in, with no re-compression.