Scanners

Printer-friendly version

Author: 

Taxonomy upgrade extras: 

I have an 80 page manuscript that I just found. It is in Typewritten print. Perhaps I did it on a Typewriter, so that would say before 82'. I also had a scanner that I could scan it in with but I have never liked the program that puts it into a file on the computer.

Has the technology improved so I could go to Kinkos or somewhere and get it done or am I going to intput it myself?

Thanks

Khaduuj

Comments

Files

It would be easy, I think. I'm not a pro. Whatever file looks like could easily be changed. Have them scan it and save it. Open your document, and highlight it for copy/paste into an Office program, or open office and then save it as unformatted or even .rtf. You can then re-open it and use the tools to inset photos, whatever. Or at least, thats the way I would try to do it.

A lot depends

On the quality of the OCR (Optical Character Recognition) software. (Our company licenses three different engines to customers, and the number of errors drops, as the price increases. The best is really good with laser printed text using standard fonts.)

Your typed document could be scanned & OCR'd (to PDF+text or MSWord, etc.) and you should expect that between 80 and 90% will be converted correctly (You might get as high as 99%).

As to whether Kinkos does it? I don't know, you'd have to ask them.

Good Luck,
Anne

Kinkos

I know the Kinkos here does it. I have a friend that works for them.

Hugs,
Jenna From FL
Moderator/Editor
TopShelf BigCloset

Hugs,
Jenna From FL
Moderator/Editor
TopShelf BigCloset
It is a long road ahead but I will finally become who I should be.

Good luck!

But bear in mind that 90% accuracy still means one error on almost every line of text.

Some of those can be fixed with a spellchecker, but you'd still have to proofread the whole document, anyway.

90% accuracy may be optimistic for typewritten text, because line spacing and character spacing, and quality of print can vary across a line, not to mention a page. One common OCR error is to mistake the combination of 'r' and 'n' for 'm', and vice versa.

Still, it may still be easier than retyping it.

Typewritten is easier than typeset

erin's picture

At least monospace fonts are easier and most typewriter faces are like Courier, monospaced.

This is Courier.

Hugs,
Erin

= Give everyone the benefit of the doubt because certainty is a fragile thing that can be shattered by one overlooked fact.

= Give everyone the benefit of the doubt because certainty is a fragile thing that can be shattered by one overlooked fact.

eeep correction -- math error

I should have said, even 99% accuracy is still almost one error per line, based on the reasoning that 99% is one error per 100 characters, and assuming 80 characters per line.

Agreed, typewritten text uses a monospaced font, but the mechanical carriage movement is not as consistent as we'd like, so the space between characters isn't consistent, which confuses some OCR programs. Add to that, variances in ribbon quality -- and of course, old-fashioned typing errors, whether they be X'd out or just missed, and 99% is a very optimistic rate.

Of course, there's a lot of variation in the quality of typewriters, too, which would make a difference.

90% accuracy is ten errors per hundred characters, that's several mistakes per line, for 80 pages. That's a lot of tedious correction. If you happen to be a decent typist, you may find it faster to retype it.

OCR

Actually, typewritten text may not be that difficult to convert, assuming a combination of decent quality typewriter, ribbon and paper (so each letter is clear, distinct, and black compared to the bright white of the paper). After all, it's all one size, one font, without any odd symbols or graphics to confuse the OCR (Optical Character Recognition) engine. Most scanners have some form of OCR software built in, and there are various commercial options available.

Scan at a fairly decent resolution (at least 300dpi) and greyscale. Save the image files, so that if the OCR program doesn't like the input, you can open the image up in a basic image editor and adjust contrast etc.

If you plan on asking a company to scan and convert for you, it might be worthwhile photocopying the manuscript beforehand, so if something goes wrong, you've still got a copy!

 


EAFOAB Episode Summaries

There are 10 kinds of people in the world - those who understand binary and those who don't...

As the right side of the brain controls the left side of the body, then only left-handers are in their right mind!

It also depends on..

It also depends on whether an electric typewriter (which has fairly uniform striking) or a manual typewriter where the force of striking the keys varies by hand and finger for most typists.

Anne

Even the inexpensive HP

Even the inexpensive HP multifunction inkjet and laser printers come with OCR and can scan a document into a pdf file
that is text not an image; i.e. you can select and copy the text from the pdf to an editor as text.

Ask friends if they have one with an automatic sheet feeder and give it a try.

I'm sure that other makes also have OCR.

Good luck

D

OCR

bobbie-c's picture

I have to agree with Anne: OCR software's pretty decent nowadays, but the best you can expect from the OCR software that comes with today's scanners is in the 90% accuracy range, and, like it was pointed by the commenters, the Word file (or PDF or TXT or whatever) it generates would have one error per ten words. But there are lots of better, commercially available stand-alone shrink-wrapped OCR programs that can convert scanned pages into OCR-translated Word files that you can edit, too, and I also believe that the improvement in accuracy goes up as the price increases as well (though I doubt that you would get 99% - as it was pointed out, typewritten text is easy to translate but the accuracy would depend on lots of quality variables).

So you would still need to proofread it.

If you don't have your own scanner or your scanner's old and therefore the OCR program's not that good, I suppose Kinko's can scan your manuscript and generate Word files for you, or you can do the OCR conversion yourself from the scan-files that Kinko's gives you. If you do, there are lots of OCR programs for sale, and if you don't want to spend, I'm sure there are lots of freeware software out there in the web that are pretty good. One of my company's clients uses a freeware OCR program and they're happy with it (I can't recommend any, though, as I haven't looked into it since my two-year-old HP scanner's OCR program's good enough for my needs).

The problem is that - there is no escaping it: you have to proofread... :(

This is all just my point of view, of course. In any case, whatever you do, good luck, and have fun while doing it.

- Bobbie

p.s. Like the fun alternate spelling of your name - "Khaduuj." Has a very Jack Black-esque sound to it.

   
bobbysig-blue2.png
To see Bobbie's stories in BCTS, click this link: http://bigclosetr.us/topshelf/book/14775/roberta-j-cabot
To see Bobbie's "Working Girl" blogs, click this link: http://bigclosetr.us/topshelf/book/19261/working-girl-blogs
To see ALL of Bobbie's blogposts, click this link: http://bigclosetr.us/topshelf/blog/bobbie-c

Thank you all, yes I'll proof read it.

Please keep in mind that this story is from the late 70's, and is such awful writing that it makes "John Norman" look good. But at least having it in a file will let me go through and edit it. The story is post destructionist, science fiction, with aliens, and not Transgender, but I can easily remedy that. I think the idea is great and is likely to become a story when I get my present work off the front burner.

Much Peace

Khaduuj