Re: [artworks] PDF possibilities
- From: Jim Lesurf <jcgl@xxxxxxxxxxxxxxx>
- To: artworks@xxxxxxxxxxxxx
- Date: Thu, 24 Nov 2016 13:05:41 +0000 (GMT)
In article <305201e455.ricp@xxxxxxxxxxxxxxxxxxxxx>, Richard Porter
<ricp@xxxxxxxxxxxxxxxx> wrote:
On 23 Nov 2016 Jim Lesurf wrote:
In article <55e3941369smartgroups@xxxxxxxxx>, Tim Hill
<smartgroups@xxxxxxxxx> wrote:
This is exactly what !JSize is for. Setting the dpi of a jpeg without
having to reprocess the image data.
There's a copy at http://timil.com/riscos
Excellent! That sounds ideal. :-)
Can't you just change the dpi in !DPlngScan and save the image without
making any other changes?
Not so far as I know. AIUI !DPingScan would render the jpeg when loading,
and then re-jpeg when saving again. Thus taking the image data though
another data-discarding process.
Whereas Tim's application seems to do exactly what is needed here. It
changes *just* the dpi values, leaving the jpeg compressed data payload
unaltered. No loss of detail.
Its similar to a problem I keep encountering. Scans sent in to us (ukhhsoc)
are often supplied as jpegs. Sometimes these are fine and can be used
as-is. But in other cases they really need manipulation.
This generates the above dilemma. The results may 'look better' if I render
them with something like !DPingScan, tweak them, and save again as a jpeg.
But in doing so some details may be lost, and the image is no longer
losslessly tracable back to the 'original'. The chain of 'provenance' and
'tracability' is weakened. So I tend to err on the side of not trying to
improve things unless I feel it really is needed.
You can find examples on the ukhhsoc website. In some cases scans are sent
in as blurred dark grey text on slightly lighter grey 'paper'. Not easy to
read. I can do the above to 'enhance' them. But this can look worse then in
some other ways, and gives rise to the above problems.
Occasionally I get situations I can illustrate if you look at
http://www.torrens.org.uk/ukhhsoc/makers/KEF/105_Active/index.html
There I got the four jpegs in the "Drawing_part_scans" subdirectory. These
were the best the sender could provide as the source material is in poor
condition and a lot bigger than A4. My best results from combining the
images (using !Compo) and tweaking the greyscale, etc, is the
105p2_Active_Circuit_Drawing_scaled_down/jpeg
file, linked at the top of the page.
Still poor. Had to alter the greyscale, blur to removed patterned streaking
of the background, etc, and scaled down to help remove surface patterning.
Any futher expansion of greyscale or gamma tweaks started to lose parts of
the drawn lines.
If the scans had been clean, the filesize would also have been much
smaller.
However the result is readable, and shows the drawing made by KEF. So has
historic value here, as well as being useful info for anyone now wanting to
study their design or service equipment.
In this case I hedged my bets by providing what I got *and* a tweaked
version. The combined result should save others having to re-do the work.
But for obvious reasons given the thousands of files involved, I tend to
want to not do this too often! :-)
When someone sends us the paper document I can work on getting a good
scan. But often we only have a submitted scan. Again, it would be difficult
to keep telling people that their scan was lousy so can they try again! I
like when possible to have a scan that looks like the source. But in the
end it is the info that matters.
Jim
--
Electronics
http://www.st-and.ac.uk/~www_pa/Scots_Guide/intro/electron.htm
Armstrong Audio
http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc
http://www.audiomisc.co.uk/index.html
------------------------------------------------------------
To change, suspend or cancel your subscription go to
//www.freelists.org/list/artworks
------------------------------------------------------------
Other related posts:
