Re: Good handling of high-res scans (avoiding thrashing)

["Martin Kretzschmar" <m_kretzschmar gmx net>]

Hi Alan,

the better performance with pdf is probably due to poppler's splash
back end (the one inherited from xpdf). It performs loading and
scaling (or even arbitrary affine transformations?) of embedded images
in one step. In cases where the image is scaled down very much, this
probably saves a lot of memory. When loading a .jpg file OTOH, the
whole image is loaded into memory (using 4961x7096x4 bytes) , and then
scaled down for display (IIRC).

Just my guess as to why pdf might perform better than raw pixbuf.

Regards,

Martin