15 Aug
2006
15 Aug
'06
3 p.m.
Barry Watzman wrote:
Re:
Agreed.
But doesn't JPEG use lossy compression?
Yes. And it blows thick industrial-waste chunks
on any
graphics that are non-photographic.
Peace... Sridhar
Bull Take a look at the stuff I've scanned on Howard
Harte's site, including
schematics, for example this one:
http://www.hartetechnologies.com/manuals/Cromemco/Cromemco%20Bytesaver%20II.
pdf
They are fine. Keep in mind that almost every PDF file that was produced by
scanning is a JPEG internally. Sure, there are lots of such files that are
crap, there are lots of bad scanners (referring to both the hardware and the
people that use them). But there are also some that are excellent, superb,
indistinguishable from the original. And since they are all JPEGs, JPEG can
do this with no discernable compromise IF YOU DON'T TRY TO OVER-COMPRESS.
I disagree with the "all scanned PDF files are jpegs", including this one.
[The schematic in the Bytesaver manual was almost
unreadable in the original
printed format, extremely fine, extremely faint lines, but take a look, for
example, at the drawing on page 10 of the manual (page 12 of the PDF file).]
Barry, there may be places here and there in the manual where things
aren't bi-level, but everything I have looked at is bilevel, including
that image. Zoom in and you'll see the characteristic stair-step on the
edges of characters.
Acrobat has smarts where even if you scan grayscale, it will classify
bilevel stuff and use G3/G4/JBIG (as per settings) or it will use a
deeper image format with none/jpeg/jpeg2000 (as per settings). Zoom in
at 2400% settings and see for yourself.
Or taking another approach -- this document has 61 pages, and the page
were 8.5" x 11" probably. The document is also 451 KB, or 7.4 KB/page,
much less than the 3.3 KB/in^2 that you recommend for jpeg. Even if you
disagree with my claims above, *something* is making this doc much
smaller than what you'd expect from jpeg encoding.
Another thing that can be confusing is that even when an image is
bilevel, acrobat will convert to grayscale on display up to some zoom
level. If you aren't zoomed in enough, things look like they have
grayscale edges, but zoom in enough and you'll see it was reconstructing
the graylevels by doing some filtering.