On Tue, 27 Jul 2004, Mats Broberg wrote:
From: ntg-context-bounces@ntg.nl [mailto:ntg-context-bounces@ntg.nl] On Behalf Of George N. White III
As a general principle, it makes no sense for pdftex to provide image manipulation capabilities. Such capabilities are useful to a much wider audience than the users of pdftex, so there are lots of tools to do image resampling and format conversions. All that pdftex should do is support inclusion of pdf. The limited support for including png images is a convenience, but if you are being careful you would want to make pdf images.
-- George N. White III
Head of St. Margarets Bay, Nova Scotia, Canada Having checked the pdfTeX documentation, doesn't the internal parameter \pdfcompresslevel deal with this? The documentation says:
"compress level This integer parameter specifies the level of text and in-line graphics compression. pdfTEX uses zip compression as provided by zlib. A value of 0 means no compression, 1 means fastest, 9 means best, 2..8 means something in between. Just set this value to 9, unless there is a good reason to do otherwise - 0 is great for testing macros that use \pdfliteral."
The compression this parameter controls is quite different. Without
compression (e.g., \pdfcompresslevel=0) a PDF file consists of almost
readable text. Even images can be stored in an ASCII encoding. Setting a
non-zero value for \pdfcompresslevel applies a lossless compression
algorithm to objects in the pdf file.
For images that will be displayed only at low resolution it may be useful
to downsample the original image to reduce the size. For example you
might have a 2 inch by 2 inch image scanned at 400 dpi. This image would
have 800x800 pixels. For screen display you might prefer to have a
200x200 pixel image (or 100 dpi for 2 inches). Downsampling refers
to the process of reducing an 800x800 pixel image to 200x200 pixels.
ways to reduce image size (lossy compression, colorspace changes,
even converting certain images to line art).
Some tools to generate PDF include methods to downsample images.
In particular, people who have been using tex-->dvipsone-->distiller
and are now using just pdftex encounter problems with much larger
pdf file sizes and excessive load times until they resize the input
images.
--
George N. White III