31 Dec
2011
31 Dec
'11
1:50 a.m.
Since several people here do this kind of thing, I figured I'd ask here.
I'm trying my hand at scanning stuff to PDF. The only thing I seem to
find for doing this in Linux is Simple Simple Scan. I'm scanning in photo
mode because text mode is unacceptably grainy. My problem now is
converting the resulting jpg into a pdf. ImageMagick is the obvious tool
for this, but no matter what I do, the resulting pdf is grainy. How can I
tell ImageMagick to not diddle with the quality?
--
David Griffith
dgriffi at cs.csubak.edu
A: Because it fouls the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
31 Dec
31 Dec
3:35 a.m.
David Griffith [dgriffi at cs.csubak.edu] wrote:
Since several people here do this kind of thing, I
figured
I'd ask here.
I'm trying my hand at scanning stuff to PDF. The only thing
I seem to
find for doing this in Linux is Simple Simple Scan. I'm scanning in
photo mode because text mode is unacceptably grainy. My problem now
is converting the resulting jpg into a pdf. ImageMagick is the
obvious tool for this, but no matter what I do, the resulting pdf is
grainy. How can I
tell ImageMagick to not diddle with the quality?
The first thing to say id that if you are scanning text JPEG is not what
you want.
You probably want TIFF, specifically G4 encoded bi-level (i.e. B&W
rather than
grey scale). If there are photographs then you might want 8-bit grey
scale
(although I've very rarely found it to make any difference, other than
to
file size).
I tend to scan @ 600dpi but anything above 400dpi seems visually
acceptable
(to me, I've no idea what anyone else thinks).
If you have colour photographs then you probably want JPEG, but only for
those
pages. For DEC manuals, I've mostly found that only the covers need
colour
scanning. There are some software manuals that have mostly black text
with the
occasional smattering of colour (or worse, blocks of text with a
background colour)
but I've never found a good way of handling those and preserving the
colour.
As for the actual scanning, I'm afraid I cannot help: my scanner
delivers the
resulting TIFF-wrapped-in-PDF via FTP :-)
Antonio
6:56 a.m.
New subject: TIFF bilevel G4, 200-400dpi - Re: scanning and conversion questions
On 31/12/11 6:35 AM, arcarlini at iee.org wrote:
David Griffith [dgriffi at cs.csubak.edu] wrote:
Good advice. Most people know by now that PDF can encode this directly too.
Since several people here do this kind of thing,
I figured
I'd ask here.
I'm trying my hand at scanning stuff to PDF. The only thing
I seem to
find for doing this in Linux is Simple Simple Scan. I'm scanning in
photo mode because text mode is unacceptably grainy. My problem now
is converting the resulting jpg into a pdf. ImageMagick is the
obvious tool for this, but no matter what I do, the resulting pdf is
grainy. How can I
tell ImageMagick to not diddle with the quality?
The first thing to say id that if you are scanning text JPEG is not what
you want.
You probably want TIFF, specifically G4 encoded bi-level (i.e. B&W
rather than
grey scale). If there are photographs then you might want 8-bit
grey
scale
(although I've very rarely found it to make any difference, other than
to
file size).
I tend to scan @ 600dpi but anything above 400dpi seems visually
acceptable
(to me, I've no idea what anyone else thinks).
I agree. Diminishing returns, 400dpi is fully adequate for text. Any
more and the resulting files become cumbersome to store and especially
read (relatively slow e-readers are becoming commonplace!)
Pay attention to diagrams with fine lines. In tricky cases, changing the
black/white threshold might be necessary to prevent dropouts. For this
reason it can be good to scan in grey scale for later processing based
on evaluating some pages (possible resolution reduction, thresholding,
and compression).
--Toby
If you have colour photographs then you probably want JPEG, but only for
those
pages. For DEC manuals, I've mostly found that only the covers need
colour
scanning. There are some software manuals that have mostly black text
with the
occasional smattering of colour (or worse, blocks of text with a
background colour)
but I've never found a good way of handling those and preserving the
colour.
As for the actual scanning, I'm afraid I cannot help: my scanner
delivers the
resulting TIFF-wrapped-in-PDF via FTP :-)
Antonio
12:33 p.m.
In article <9586467B685849F792CA0C405FECF717 at ANTONIOPC>,
<arcarlini at iee.org> writes:
grey scale). If there are photographs then you might
want 8-bit grey
scale
(although I've very rarely found it to make any difference, other than
to
file size).
In my experience, scanning in grayscale or color for photographs is
essential to making the photographs useful. On the other hand,
scanning in B&W is essential for good text representation for OCR, so
if you're looking at documents that mix lots of text and photos,
you're in a quandary about which to do. What I do is scan such pages
twice. My intention is to have a PDF that layers the grayscale photo
on top of the B&W text page, but I haven't figured out the best way to
do that yet.
--
"The Direct3D Graphics Pipeline" -- DirectX 9 version available for download
<http://legalizeadulthood.wordpress.com/the-direct3d-graphics-pipeline/>
Legalize Adulthood! <http://legalizeadulthood.wordpress.com>
7:42 a.m.
On Dec 31, 2011, at 4:50 AM, David Griffith wrote:
Since several people here do this kind of thing, I
figured I'd ask here.
I'm trying my hand at scanning stuff to PDF. The only thing I seem to find for doing
this in Linux is Simple Simple Scan. I'm scanning in photo mode because text mode is
unacceptably grainy. My problem now is converting the resulting jpg into a pdf.
ImageMagick is the obvious tool for this, but no matter what I do, the resulting pdf is
grainy. How can I tell ImageMagick to not diddle with the quality?
--
David Griffith
There are some notes about this on the home page of bitsavers.org. Among other things
they say:
Post-processing is done using Lemkesoft's Graphic Converter
TIFF to PDF conversion is done using Eric Smith's tumble
The preferred form for any contributed text scan is as a collection of lossless
Group 4 fax compression (ITU-T recommendation T.6) images saved as TIFF
files with a minium scan resolution of 400 dpi.
Lower scan resolutions produce noticable artifacts if a page needs to be
straightened in post-processing.
Lossy compression formats, such as JPEG, should NEVER be used to save pages
of text, since the compression format destroys edge resolution and contrast
would make it difficult to OCR in the future.
-Jeff
9:56 a.m.
On Sat, Dec 31, 2011 at 01:50:58AM -0800, David Griffith wrote:
Since several people here do this kind of thing, I figured I'd ask here.
I'm trying my hand at scanning stuff to PDF. The only thing I seem
to find for doing this in Linux is Simple Simple Scan. I'm scanning
in photo mode because text mode is unacceptably grainy. My problem
now is converting the resulting jpg into a pdf. ImageMagick is the
obvious tool for this, but no matter what I do, the resulting pdf is
grainy. How can I tell ImageMagick to not diddle with the quality?
For my archive, the scan workflow is:
- use scanimage to get a (set of) 600 dpi bilevel (monochrome) scan(s) at
A4 size
- deskew the scans using tiff_findskew, gm convert and pnmrotate
- convert tiff(s) to Fax G4 compression
- build a tar archive of the tiff(s) in case I need it later (e.g. OCR)
- use tumble to produce a (multipage) pdf from the tiff file(s)
- use cjb2 and djvm to create djvu a file
- dump the resulting tar, pdf and djvu files into a git repository
that is replicated over several machines (and also goes out to tape)
The above has, of course, been scripted appropriately since I'm lazy ;-)
Scripts available on request.
Kind regards,
Alex.
--
"Opportunity is missed by most people because it is dressed in overalls and
looks like work." -- Thomas A. Edison
10:58 a.m.
I feel the loss of detail by scanning bilevel can throw information
away, this is especially important with
circuit diagrams and images, while automation may speed up the work
some attention to detail is required else the end product will be less
than useful. I currently stay with manual scanning and greyscale . and
dont squeeze a quart into a pint pot.
As for PDF and and slow memory hog PDF readers/display programs I
think it may be time to wean users away asap.
I have recently tried diva as an online display tool
the following manual is incomplete, all I have, source is a poor
photocopy but the diagrams are just about readable
http://www.collection.archivist.info/diva/systrondonner1626.html
direct link to part of the diagram
http://www.collection.archivist.info/diva/systrondonner1626.html#p=138&…
another complete with colour
http://www.collection.archivist.info/diva/lucastp1.html
The system uses tiled tiff
https://github.com/DDMAL/diva.js/wiki
Main advantage is no need to download a full pdf, light load on the
server, fast viewing.
can give direct links to a page and zoom
option can have two panels to view setup and diagram at the same time.
btw the above demos are hosted on an ADSL line 7.65meg down .35meg up
(long way to the exchange)
Dave Caroline
4524
days inactive
4524
days old
6 comments
7 participants
participants (7)
-
als＠thangorodrim.de -
arcarlini＠iee.org -
dave.thearchivist＠gmail.com -
dgriffi＠cs.csubak.edu -
jeff.ogden＠umich.edu -
legalize＠xmission.com -
toby＠telegraphics.com.au