Scanning electronic magazines

May 01, 2006 17 Replies

Hi All.



Trying to de-clutter my life by scanning hoarded electronics magazines to CD ROM.



However, I find that much material uses 'halftone' techniques to achieve greyscale shading. Looks fine to the human eye but horrible when scanned. All sorts of Moiré patterns appear, and the image files are huge. A simple halftone grey background generates a lot of black and white pixels.



Is there any software that can have a good go at reversing the halftoning process, i.e. convert halftone images back into greyscale?



I reckon it is possible, I just have not seen any suitable software around.



Google just seems to return mostly academic studies on the subject.


What display means do you have to show a gray image? It will, of course, not involve pixels.

Many thanks, Don Lancaster voice phone: (928)428-4073 Synergetics 3860 West First Street Box 809 Thatcher, AZ 85552 rss: http://www.tinaja.com/whtnu.xml email: don@tinaja.com Please visit my GURU\'s LAIR web site at http://www.tinaja.com

What program are you currently useing to scan with? You can set on most programs to scan in at like 150 dpi and grayscale.

I have a lot of old mags, if you need more input to your collection. Jtt

My Canon scanner (not that I'd necessarily recommend it) has a 'magazine' mode that seems to work okay on such material. I scanned a ~40 page color offset litho catalog to have a copy with me on a trip and it came out to about a 20M PDF. That's with color retained.

Best regards, Spehro Pefhany

"it\'s the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com

Hello Spehro,

Which brings up a point: What would be the minimums in resolution and color/gray scale retainment required for a "paperless office"? Minimums as to agency requirements, IRS and others. Several times I tried to find official guidelines but there wasn't much. It would be nice if at least some stuff could go electronic.

Regards, Joerg

formatting link

I've got about half my articles digitized.

For monochrome 1-bit images, I've found 300 dots-per-inch is fine--as good as an excellent fax--*except* you'll lose the occasional subscript, superscript, or fine-print. Those items become too pixellated to read at that resolution. 600 dpi binary images are great, tiny details are clear.

100dpi 8-bit grayscale is surprisingly readable, and is what I've stored some older magazine articles with photos in, but that's max'd out: 'zooming' in reveals no further detail. It's much nicer to use 200+ dpi in grayscale, and then be able to zoom in and see details clearly when things aren't so clear.

Optical character recognition (OCR): many programs work MUCH better with 600dpi 1-bit images (e.g. Pagis Pro), though my ancient version of PaperPort likes 300dpi fine, and recognizes more accurately than its (former) competitor with 4x the pixels.

HTH, James Arthur

The moire patterns are a result of the interaction between the dot pitch of the halftone images and the dotpitch of your scanner.

I find a good way to get rid of them is to scan images at a higher resolution than you ideally want ( try 300 dpi as a starting point ), then use something like Paint Shop Pro to blur the image. Once blurred, reduce the image resolution, e.g. to 50% of that scanned in.

You may need to play with the initial scan resolution, and the particular blurring, or averaging algorithm, but you should be able to get something that works well.

I have a recollection that the freeware IrfanView imaging program may include a mechanism for automating the process, so that may be worth a look.

CD

ple

d=2E

Kryten, All the scanning utilities I've used include an option to handle half-tone images, often a scan-time check-box labelled "descreen." It usually works very well.

An option of last resort is to scan photos and illustrations in grayscale, which works well, but makes for large scan files.

I've sometimes scanned the same article twice: once in 600dpi 1-bit mode to ensure reliable OCR, and once in 100-150 dpi 256-level grayscale to get any pictures.

HTH, James Arthur

yellow.yeti @ gmail.com wrote WITHOUT CONTEXT:

Re: Google Groups, "context", and your posting style:

formatting link
't-see-*-*-*-*-Google+You-don't-*-*-*-*-*-*-*-previous-post+click-*-show-options-*+zz-zz+*-leave-enough+click-THAT-Reply-link+If-you-don't-*-*-*-*-your-reply+faqs http://66.102.9.104/search?q=cache:SXIajvWUVHAJ:groups.google.com/support/bin/answer.py=answer=12348+Summarize-what-*-*-*-*-*+remove-*-*-*-irrelevant+show-options+zz-zz+click-Reply+BOTTOM+*-*-reply-appears-*-*-*-before-the-original-*-*

Hello James,

That's pretty much my experience as well. I just don't know what an agency such as the IRS would accept for documents where the originality of a signature would not be an issue or where there is no signature. Receipts and stuff like that. Is there any official guideline?

Paperwork reduction act and all that sounds nice but only with clear guidelines will it make any sense.

What really surprises me is the clarity of bitmaps imported into MS-Word docs. Clients usually want docs in that format. Yesterday I imported a pretty busy letter A size schematic in 300dpi monochrome bitmap into Word. It contains lots of non-rectangular graphics like logic gates in US standard notation. Word still shrunk the file size of that portion down to about 200k. I always print out both the CAD sheets and the document and usually cannot see a difference in quality when printed our letter size.

Regards, Joerg

formatting link

Hello James,

IIRC my first scanner came with an overlay filter. It was yellowish and I don't know how it worked but it caused grayscales to become less distinct. This hand scanner couldn't do grayscales so one had to get rid of them as much as possible before the optics eyed the document.

Regards, Joerg

formatting link

Thanks to all for the comments.

Joerg asked about resolutions. IIRC the PC's ergonomics assumes about 72 dpi, and I reckon this is practical. My screen is 12x9 inch, and 1400x1050, thus about 117 dpi which strains my eyes a bit. I'd rather it were a 24" diagonal screen :-)

I have used 300 dpi to scan black and white text. This renders the characters well enough for OCR.

600dpi doesn't give noticeably better recognition.

With halftone images, it has been suggested that these should be scanned at least four times the frequency resolution of the halftone image in order to be of adequate quality.

I'm surprised halftone input isn't a common feature in scanner software. There must be quite a lot of halftone printed material, needing to be scanned.

Maybe it is just a hard task for software to do an intelligent job of it.

Hello Kryten,

My question was more about regulations, as to what would be acceptable for certain categories of documents if you ever were audited.

The monitor on my CAD PC is huge but even there 150dpi and 300dpi non-zoomed doesn't make much of a difference.

Regards, Joerg

formatting link

Howdy Joerg, I suspect the yellow filter was to provide your scanner with monochromatic input. The "descreen" option I mention is a software filtering option, no changes to the optical input are required.

I believe the "descreen" option my scanners` drivers offer (I have rather a collection of them) simply automates the process Yellow Yeti described above: oversampling at high resolution, then filtering and re-sampling. The results are usually quite good. Images riddled with Moire patterns clean right up.

Regards, James

No, not really. As long as the electronic copy accurately reflects the content ofthe original, i.e. no electronic editing monkey-business allowed. Faxes are legal, if that helps.

Not the Paperwork reduction act, it is the Electronic Signatures Act (Public Law No: 106-229) of 2000 that applies. Sections 101(d) and

104(b)(3) are the only relevant sections I see regarding required image quality, and they don't specify any standard of resolution or clarity:

formatting link
Regards, James

Hello James,

Well, there should be :-)

That does help because it would indicate that the usual fax resolution might suffice.

I'll have to try that tomorrow. The search has timed out and when I re-entered the title it came up with >1000 hits.

Regards, Joerg

formatting link

Sorry Joerg, I didn't notice it was a timed link. This should work better:

S.761 Electronic Signatures in Global and National Commerce Act

formatting link

Best, James Arthur

Hello James,

Thanks. Looks they they put a little chicken exit in there unless I misunderstand something: " Notwithstanding any statute, regulation, or other rule of law (other than this title and title II)...".

So an agency could possibly say "Well, that may all be fine and dandy but we have recently issued a little regulation that we will not allow...".

Regards, Joerg

formatting link

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required