There are two kinds of PDF's. One is an image (with or without some search text) and the second is pure text which is likely the kind you are testing with. Typically DJVU shines with an set of images that look like text. DJVU looks at the images and breaks them up into reusable graphics items that look for all intents and purposes like a set of fonts and then just references these as needed. This is typically better compression than an image would be, also the latest PDF is now using a better (JP2000) compression than it did formerly.
Dale
|