[TheForge] OCR

[Roy Wilson]

On Sun, 3 Feb 2002 12:22:56 -0600, lama wrote:

>Daniel, thanks for your very generous and kind offer.
>If we average 64 pages x 4 issues a year x 25 years,,,,, = 6400
>pages for the Anvil's Ring x ( how many?) for Hammer's Blow
>we are looking at what, 8,000 pages more or less. Scanning cost at least
>a buck a page (How much Roy Wilson?) 

	I started this job.  The Anvil's Ring did *not* scan well. 
Terrible recognition from the scans, the combination of gloss
paper, odd font (letters that touch each other), and the paper
thickness gave me no end of trouble.  The files were huge, even
when OCR'd and saved in Word (for easy conversion to HTML).  The
pics need to be scanned alone at a much higher resolution and then
stepped down (my 58 page/minute scanner for OCR only goes up to
300dpi).

	I finally ran the two issues off by scanning direct to PDF
(huge!), burned them to a CD and sent it to ABANA.  Never heard
back on it, I assumed they felt it was unusable on a website (and I
agree).

	Daniel - if you're willing to do it, you really need at
least 600dpi to capture the pics as well as the text.

	And since you're doing so much OCR, contact me offlist
<rwilson9@twcny.rr.com>, I've found an untouched venue for OCR
services that I'm willing to share - no way I can handle the amount
of it that even my county produces, much less any other county or
state.



Roy Wilson
General Operations Director
Zanzibar Internet Land Line Administrator