AKICIDW [*]

[piranha]

with the boat work really entering the last stages now, i'm looking at my shelves full of books, and ponder scanning them. now, i have lots of experience scanning manga, as well as scanning prose and OCRing it. but the books i most want to scan myself because there are no digital copies out there are craft books.

which are often beautifuly laid out, and that presents a problem. of course the easiest is just making a pdf from the images, but that results in enormous files, and means the text can't be be searched. i'm big on searching these days, especially when it comes to patterns.

does anyone have experience with scanning such books and recreating something similar to the original before it went to press? any suggestions for software (i prefer mac, but can deal with windows and linux)? workflow?

[* = all knowledge is contained in dreamwith]

Comments

[serene]

They do this at my job (that is, scan books of all sorts for students with visual and other disabilities) and I will ask Martha, who is in charge of that, what she recommends.

[piranha]

thank you!

[jesse_the_k]

I have both composed (typesetter) and decomposed (Braille transcriber) such works. Sometimes the paper formats are the best possible combination: if you can use the paper books, I'd recommend moving to a different task. The craft books on my shelves have text that relies on an image in close proximity.

[piranha]

thanks for the info -- don't really have a choice, unless i want to do without the books. i might just stick with the scanned images if it is a total pain. for some of them there will eventually be digital copies i am sure, but never for the older ones.

[jesse_the_k]

Must have books!

Upon further reflection, scans will be OK once we routinely have monitors which permit 100% viewing of an 8x10 in page. I get many beadworking patterns as downloads these days, but I have to print them out so I can see all the info at once.

[piranha]

i love japanese crochet books -- they hardly use any text, but symbol charts. together with a good picture, i find those much more easy to follow than long descriptions of row 1 through row 92.

adobe acrobat seems to do a so-so job of sorting text from images, which is making me hope this isn't as much work as i'm thinking right now. except acrobat and i, we're not on good terms. it's the least intuitive software i've ever worked with.

[juliet]

Decompose into something resembling LaTeX source format? Where pictures are put in with a label tag & a sourcefile tag (& then you'd have the actual image file saved elsewhere), & you use the label to refer to them in the text.

e.g.

blah blah blah see Figure \ref{laceknitting1}.

\label{laceknitting1}{\figure{laceknitting1.jpg}}

Sorry, that's is not accurate LaTeX source code -- I haven't written LaTeX in a while & can't be bothered looking it up, I'm afraid :) But you get the idea!

[piranha]

hmmm. i hadn't thought along those lines at all -- good idea, thanks!

[gam0ra.blogspot.com]

Is there some reason the text can't be searched in pdfs the way you make them? I often search pdfs for text contained within them, but I know nothing about making pdfs.

[piranha]

all sorts of things can be inside a PDF; text (which is searchable) is only one of those. it can also contain drawings (vector art) which can be resized easily. or raster images (which cannot be resized).

when one scans, one creates a raster image of text and/or graphics. the text is not machine-readable unless it has been through a process called OCR (optical character recognition). OCR is pretty good these days, if you make a good, clean scan. the problem is that craft and other how-to books contain lots of images lined up with text that refers to them. which means automatic processing (shove files in one side, get searchable text out the other) goes kind of out of the window.