Scanner recommendation

[Greg Anos]

As an experienced scanner, you don't want direct to HTML. Using a word processor format is much better to start with because you can then use the spell checker for your scan, which is awkward with an HTML file. Then use the word processor to do the HTML conversion.

[ascherjim]

Originally Posted by Ralph Sir Edward

As an experienced scanner, you don't want direct to HTML. Using a word processor format is much better to start with because you can then use the spell checker for your scan, which is awkward with an HTML file. Then use the word processor to do the HTML conversion.

I couldn't agree more. I fear that beginners to ebook creation through scanning underestimate the amount of editing (proof-reading) required, even when having scanned the most legible printed pages. And the use of word processors in the process, with all the editing tools (especially global) at their disposal, tend to simplify and shorten (somewhat) and relieve the tedium (somewhat) of the process.

[Skydog]

Originally Posted by Ralph Sir Edward

As an experienced scanner, you don't want direct to HTML. Using a word processor format is much better to start with because you can then use the spell checker for your scan, which is awkward with an HTML file. Then use the word processor to do the HTML conversion.

Indeed, but aren't most word processors, particularly Word, known for terrible HTML output?

[texasnightowl]

Originally Posted by Skydog

Indeed, but aren't most word processors, particularly Word, known for terrible HTML output?

Word is really really bad unless you choose the "filtered" html output...then it is better...at least comparatively speaking.

[ascherjim]

Originally Posted by texasnightowl

Word is really really bad unless you choose the "filtered" html output...then it is better...at least comparatively speaking.

My personal methodology -- which gives me ultimately my ebook in my preferred Mobipocket format -- is so convoluted that I am loathe to really recommend it to anyone else. I scan to txt, convert that to Wordperfect for editing, then convert it back to txt for conversion to html using txt2html, then convert from html to Mobipocket using Calibre. Crazy as it sounds (and less than perfect as it occasionally is) it works fine overall for me!!!

[Skydog]

Originally Posted by ascherjim

My personal methodology -- which gives me ultimately my ebook in my preferred Mobipocket format -- is so convoluted that I am loathe to really recommend it to anyone else. I scan to txt, convert that to Wordperfect for editing, then convert it back to txt for conversion to html using txt2html, then convert from html to Mobipocket using Calibre. Crazy as it sounds (and less than perfect as it occasionally is) it works fine overall for me!!!

Actually, it does not sound crazy at all - I follow your logic! I'm willing to try anything that will result in success.

[ascherjim]

Originally Posted by Skydog

Actually, it does not sound crazy at all - I follow your logic! I'm willing to try anything that will result in success.

Of course, before I settled on my system I tried (experimented with) many other options, which I encourage everyone to do. What gives me what I want in the end is not necessarily what would give anyone else what they might want. It depends on so many factors. Patience and perseverance are the requisites.

[Greg Anos]

Originally Posted by Skydog

Indeed, but aren't most word processors, particularly Word, known for terrible HTML output?

First, let us define terrible. Seriously. I define bad two ways, One, it doesn't function. Two, other programs I need to read the HTML won't read it.

I don't insist on pretty, tight HTML, just does it work, and can it be converted to Epub and Mobi, and be read in FBReader from a zipped file. (I use OpenInkPot on my Hanlin readers....

Now this is a very loose standard, inasmuch as I will go in and add my own bits of HTML to the output. (Usually images and internal hyperlinks.)

Now by this standard, Word 97 works somewhat (won't convert into anything in Calibre, but will convert to epub in Sigil and the Epub can be back converted into mobi by Calibre), Word 2000 works well, but produces XML (XHTML) which is needlessly wordy for e-book use compared to HTML, Atlantis doesn't work (can't handle RTF conversion properly and sets some lines of text to zero height.), and Open Office messes up the RTF formatting while editing it, which leads to bad HTML formatting output.

So currently I use scanner to RTF, spell-check in Open Office despite it's problems, because it highlights hypenated words nicely, clean up the format mangles in Wordpad, load into Word 2000 and save as web page, and load into Calibre, which zips up the XML (and converts to Epub and Mobi, if required.)

Oh yes, I also use Hexedit 3.0 to clean up RTF control language at byte level, as needed.

Anyway you look at it, lots of work...

[pomo81]

You can imagine my joy in finding this forum. I've been scanning texts for archival purposes for nearly 20 years. I've spent the past two days scouring this forum and getting introduced to fantastic software I had not known about. Abby, Omnipage, and Paperport I'm familiar with; Scan Tailor and ScanKromsator are new to me and had been desperately needed.

I don't have much need for OCR, but fast flatbed scanning is a must. In nearly 20 years, I have not come across a flatbed scanner that is faster than the old -- and big and heavy -- HP 4C scanners, which can do 4 300 dpi 1-bit (black and white) legal-sized scans in one minute. Opening up a book on the glass and doing two facing pages at a time will yield 240 scans (480 pages) an hour.

All of the "fast document" scanners seem to have a speed rating only on the ADF, not the flatbed, which seem to be terribly slow. I think some of the Epson flatbed scanners might be able to come close to 4 legal-sized 300 dpi, 1-bitscans a minute, but I'm not certain. The old AGFA SCSI scanner did 3 scans a minute.

I have no idea whether I can get my HP 4C, which can still be bought new or in very good condition on Ebay for nearly nothing (it's the shipping that costs), and required SCSI card to work on a Vista or Windows 7 machine, as I have mine connected to an XP machine.

I'm always on the lookout for a faster flatbed USB 2.0 scanner, but I've seen nothing. I'm thinking about a digital copier that can be connected to a PC, as I'd like the flatbed scanning time to be just as fast as a copier. I'm also thinking of going the route of a photo box with two SLR cameras.

Anyway, I just wanted to throw in a word on these old 4C scanners and finally take part in this forum, which I had sorely needed the past 20 years and found only two days ago.

[Kirtai]

There's a forum about building book scanners based on digital cameras at http://www.diybookscanner.org/ you might find interesting. I believe the ScanTailor developer posts there too.