[sane-devel] Sugg: frontend to scan and append to pdf

Jonathan Buzzard jonathan@buzzard.me.uk
Fri, 12 Dec 2003 17:24:06 +0000


anoah@pfeiffer.edu said:
> the problem is the freaking huge filesizes you will get. perhaps you
> should instead look at one of the multi-image in single file formats,
> like  tiff. are you doing bw, gray, or color scans?

Depends on the page. If the page has no colour on it I do a greyscale
scan and convert to B&W later. If it has colour on it I do a colour
scan.

Personally I always put the images through a graphics editor for
tidying up. I am usually doing it because the page in question is
very old and tatty, so there may well be considerable reconstruction
going on.

If you are using a single image for a page which is basically all text
with a single small grayscale or colour graphic, then you are doing it
wrong. You need to separate the page into different graphics for the
different bits. A single BW 300dpi PNG as a backdrop containing all the
text, then a number of over layed JPEG's containing *just* the graphic.

In this manner a 29 page manual with some 6 grayscale graphics is a
just 2.4MB rather than 14MB if I had done the whole lot as 29 300dpi
grayscale JPEG's. My problem is all the examples I have are copyrighted
so I cannot easily show you them. For 

There is a bit of windows software that does this automatically,
but I cannot find the link now for trying.


JAB.

-- 
Jonathan A. Buzzard                 Email: jonathan (at) buzzard.me.uk
Northumberland, United Kingdom.       Tel: +44 1661-832195