GIGI questions

[Jules Richardson]

On Wed, 2005-05-11 at 07:55 +0200, als at thangorodrim.de wrote:
> > [imagemagick]
> > Note that it's *not* quick by any means and eats memory; Imagemagick
> > buffers everything as 32bit internally (the one thing I hate about it!)
> > so even if you have mono images as input it's still going to treat the
> > images as 32bit before appending to the output pdf file.
> 
> Yes, it is a _serious_ memory hog and quite slow too.

To be fair, it's seriously quick on large images with multiple bit
depths though (say, digital camera sorts of resolutions and above). It's
probably efficient on memory usage for those types of images, although
I've never checked...

We used to regularly chuck 10k x 10k pixel images at it and apply
various transforms, and it was *way* quicker than anything else we could
find on the market (free or otherwise) at the time. 

Hopefully one day they'll modify the code so that images with 8bpp or
less are handled more efficiently for certain operations (probably
controlled via a command-line flag)

> > Funnily enough  the first thing I do when I download a bunch of scans in
> > PDF format is convert them to multiple TIFF images, as they're easier to
> > handle in whatever app is appropriate for what I'm trying to do versus
> > some sucky PDF viewer ;-)
> 
> Just for reading a scanned article I find the PDF version easier to
> handle. But still keep the original TIFF files just in case.

Each to their own :)  Usually I'll just use my favourite thumbnail
viewer to browse / read - I just find it handy to be able to quickly
deal with individual pages rather than something encapsulated in another
file I suppose. Plus of course I use Linux on the desktop and all the
Linux PDF viewers are crap :-)

cheers

Jules