Show whole topic
Apr 22, 2008 11:50 am
Apr 22, 2008 11:50 am
Subject: Re: Creating a Rapier knowledge base on line?
Hi,
Sounds an interesting project though I'm not a Rapier owner (LG45).
I'd be surprised if you need several gigabyte of space - that'd be thousands of images at quite a large size. I'm a graphic designer by profession and maintain a few websites which are image heavy. As a rough guide 500 images at 600 x 460 pixels would take up around 120 mb or so. And that's at a high jpeg setting (fussy clients). Less if jpegged down a bit. You save space too by making sure black and white photos are scanned as greyscale and not RGB (colour). If the images are big you use a lot of bandwidth as well as storage space.
Text takes up much less space than an image so OCR scanning is the best way to do text. You do need to proof read it though as, for instance, an "r" and a "n" close together can turn into an "m" etc. Or at least run a spell checker. You do need to get the original flat on the scanner which can be a problem if in a book. A blurred edge of a photo scanned from a book may be acceptable but if it's text you end up having to key in the missing bits as the software won't read the bit you can't get flat. So really depends what your originals are whether OCR is practical.
Also don't ignore copyright issues.
David Eldred
Hi,
Sounds an interesting project though I'm not a Rapier owner (LG45).
I'd be surprised if you need several gigabyte of space - that'd be thousands of images at quite a large size. I'm a graphic designer by profession and maintain a few websites which are image heavy. As a rough guide 500 images at 600 x 460 pixels would take up around 120 mb or so. And that's at a high jpeg setting (fussy clients). Less if jpegged down a bit. You save space too by making sure black and white photos are scanned as greyscale and not RGB (colour). If the images are big you use a lot of bandwidth as well as storage space.
Text takes up much less space than an image so OCR scanning is the best way to do text. You do need to proof read it though as, for instance, an "r" and a "n" close together can turn into an "m" etc. Or at least run a spell checker. You do need to get the original flat on the scanner which can be a problem if in a book. A blurred edge of a photo scanned from a book may be acceptable but if it's text you end up having to key in the missing bits as the software won't read the bit you can't get flat. So really depends what your originals are whether OCR is practical.
Also don't ignore copyright issues.
David Eldred